Asset 20 8 2
Does AI recommend your business? Run the free check →

Join 15,000 business owners, marketers and entrepreneurs. The Sunday newsletter you'll be annoyed only arrives once a week.

Article

Midjourney Prompts for Fantasy Art That Work

If you are skim reading
The short version: good Midjourney fantasy prompts are built from five layers, subject, style reference, lighting, composition and technical parameters, in that order, and the biggest mistake people make is skipping the parameters and wondering why their "epic

The short version: good Midjourney fantasy prompts are built from five layers, subject, style reference, lighting, composition and technical parameters, in that order, and the biggest mistake people make is skipping the parameters and wondering why their "epic fantasy battle" looks like every other epic fantasy battle on the internet. I'll show you the exact structure, a worked example you can copy, and why most fantasy art prompt lists you've read are quietly useless.

Why most fantasy prompts you find online don't work

Search "Midjourney fantasy prompts" and you'll find hundreds of lists that are just adjectives stacked on top of each other. "Epic, cinematic, detailed, intricate, masterpiece, 8k, fantasy." None of that is a prompt. It's a mood board written in single words, and Midjourney treats it as a mood board too, which is why you get generic results that look like stock art from a game nobody's heard of.

The uncomfortable bit nobody says out loud: Midjourney's fantasy outputs are converging. Because millions of people are using the same thirty buzzwords (ethereal, mystical, enchanted, otherworldly), the model has learned a kind of average fantasy look, and that average look is exactly what you get when you prompt lazily. If your goal is art that doesn't look like everyone else's Discord feed, the buzzwords are the problem, not the solution.

The five-layer structure that changes your output

I build fantasy prompts in this order, every time, and it's the single biggest improvement you can make without touching a single setting:

  • Subject, specific, not generic: not "a wizard" but "an elderly hedge wizard mending a torn map by candlelight"
  • Style reference, a named artist, movement or medium: "in the style of Brian Froud" or "gouache illustration, 1970s children's book"
  • Lighting, named, not just "dramatic": "single candle light source, long shadows, warm amber tones"
  • Composition, where the camera is: "wide shot, low angle, subject in lower third"
  • Technical parameters, aspect ratio, stylisation, version: --ar 16:9 --s 250 --v 6.1

Drop any one of these five and Midjourney fills the gap with the average. Add all five and you get something specific enough to look like an image rather than a vibe.

A worked example

Say you run a small tabletop games company and you need cover art for a homebrew bestiary. You want a forest guardian creature, not a dragon, something original. Here's how the layers build:

  • Weak prompt: "mystical forest guardian, epic fantasy, detailed, 8k"
  • Layered prompt: "a moss-covered antlered guardian the height of an oak, standing in a flooded birch forest, bioluminescent fungi clinging to its antlers, in the style of Ivan Bilibin folk illustration, dappled green light through canopy, low angle wide shot, creature centred, --ar 3:2 --s 400 --v 6.1"

Run both through Midjourney and the difference isn't subtle. The weak prompt gives you a generic green humanoid. The layered one gives you a specific creature with a specific silhouette, specific light, and a specific illustration tradition behind it, something you could put on a book cover without it looking borrowed.

Parameters worth knowing for fantasy work specifically

A few settings matter more for fantasy art than for other genres:

  • --s (stylise) controls how much artistic licence Midjourney takes. For fantasy, push it to 250 to 500. Lower and you get literal, flat results. Higher and the model starts painting rather than illustrating, which suits fantasy's more painterly tradition.
  • --chaos between 10 and 30 is useful when you're generating concept variations for a character and don't want four near-identical grid results.
  • --ar matters more than people think: 2:3 for a character portrait, 16:9 for a landscape or battle scene, 1:1 if it's going on social media as a square post.
  • --v 6.1 handles hands, armour detail and crowd scenes noticeably better than older versions, which matters a lot for battle art where limbs used to turn into noise.

Named artist references that do something

Vague style words ("fantasy style", "magical") do almost nothing. Named references do a lot, because the model has actual training data tied to those names. A few that reliably shift the look of fantasy output:

  • Brian Froud for whimsical, earthy, folkloric creatures
  • Frank Frazetta for muscular, high-contrast heroic fantasy painting
  • Yoshitaka Amano for elongated, delicate, ink-wash character design
  • Ivan Bilibin for Slavic folk illustration with flat decorative borders
  • John Howe for architectural, Tolkien-adjacent landscapes and castles

Pick one, not four. Stacking four artist names doesn't blend them neatly, it muddies the output because the model is trying to average four very different visual languages at once.

Character consistency across a series

This is the part that trips people up most. If you're building a character for a story, game or comic and need them to look the same across ten images, a single text prompt won't do it reliably. Use Midjourney's image prompting by uploading a reference image at the start of your prompt string, then keep your text description of the character identical in every subsequent prompt, changing only the scene around them. Expect some drift anyway, Midjourney isn't built for frame-to-frame consistency the way a dedicated character tool is, so budget for regenerating and picking the best match rather than expecting a perfect match every time. If you're building something closer to sequential panels rather than single images, it's worth looking at free tools built specifically for comics and manga rather than forcing Midjourney to do a job it wasn't designed for.

Lighting and mood words that earn their place

Skip "atmospheric" and "moody" on their own. Use these instead, because they describe a physical light source Midjourney can render:

Work with me

Want AI doing the heavy lifting in your marketing?

I build the systems that handle the boring 80 percent, so you get your week back. Done properly, with the human kept in.

  • "torchlight from below, upward shadows on the face"
  • "cold blue moonlight through stained glass"
  • "golden hour backlight, rim lighting on armour edges"
  • "bioluminescent glow as the only light source"

Each one tells the model where the light is coming from and what it's doing, which is the actual job of a lighting description. "Moody" tells it nothing.

Where this overlaps with other AI image work

Fantasy prompting isn't a separate skill from the rest of AI image generation, it's the same five-layer logic applied to a different subject. The composition and lighting principles that work for winter portrait prompts or seasonal card art are the same ones that make a forest guardian look deliberate rather than default. If you're producing commercial images rather than personal art, the same discipline applies that I'd use for product photography prompts: specific subject, named style, named light, named composition, every time.

If you want to compare tools rather than commit to Midjourney specifically, it's worth reading up on what Perchance's image generator does before you pay for a subscription you might not need for casual fantasy doodling.

A quick checklist before you hit generate

  • Is the subject specific enough that two different people would picture roughly the same thing?
  • Have you named a style reference instead of describing a mood?
  • Have you named a light source, not just an adjective?
  • Have you set an aspect ratio that matches where the image is going?
  • Have you avoided stacking more than one artist name?

If you can tick all five, you've already beaten most of what's generated in Midjourney's Discord on any given day.

This guide is part of my ChatGPT Prompts and Limits: 30 Guides to Fun Prompts, Photo Edits and Daily Caps. To write your own, try the free ChatGPT prompt generator.

Frequently asked questions

What's the best Midjourney version for fantasy art?

Version 6.1 is currently the strongest for fantasy work because it handles hands, armour detail and crowd scenes with far less distortion than earlier versions, which matters a lot for battle scenes and detailed character art.

How do I stop my fantasy art looking generic?

Drop the stacked adjectives like "epic, mystical, detailed" and replace them with a specific subject, a named artist style, a named light source and a set composition. Generic prompts produce the model's statistical average, and that average is what millions of other users are already generating.

Can Midjourney keep a fantasy character consistent across images?

Partially. Uploading a reference image alongside a consistent text description helps, but expect some drift between generations and plan to regenerate and select the closest match rather than expecting identical results every time.

What stylise setting works best for fantasy art?

Somewhere between --s 250 and --s 500 tends to suit fantasy's painterly tradition, giving Midjourney enough artistic licence without losing the specific details you put in the prompt.

Run a business, a brand or a blog? I feature tools, products and expert guest posts for an audience of 15,000. See how it works

Published and maintained by the Lilach Bullock team, covering marketing, AI and business growth.
Your buyers are asking AI who to use. Does it say you?

See for free whether ChatGPT, Claude, Perplexity, Gemini and Google name you, and get the plan to become the answer.

Check my AI visibility →
Sundays only

Get the Sunday newsletter.

One email a week. AI experiments, marketing tactics, and the workflows Lilach is building right now in her own business.

Subscribe free

Let’s get your marketing running on AI.

Book a free 30-minute call

We figure out what you need, where AI fits in, and what working together would look like.

Book the call →

Or take the 30-second calculator

You’ll see the hours and the money quietly leaking out of your week, and the three workflows worth building first.

Take the calculator →

Or grab the free AI resource library

Prompt packs, templates, checklists, and swipe files. The exact tools I build for paying clients. Yours, free.

Get the library →
Keep reading

More from the blog.