Midjourney Prompt Formula for AI Art

Create clearer AI art directions with these structured Midjourney prompts for portraits, landscapes, and abstract art.

The Prompt

Create Midjourney prompts using this advanced formula:

[Style] [Subject] [Environment] [Lighting] [Color Palette] [Mood] [Composition] [Technical Parameters]

Examples:
1. Portraits: "Cinematic portrait of [subject], dramatic Rembrandt lighting, golden hour, shallow depth of field --ar 2:3 --q 2 --v 5.2"

2. Landscapes: "Epic fantasy landscape, floating islands, bioluminescent plants, volumetric fog, sunset colors, warm hand-painted animation look --ar 16:9 --q 2"

3. Abstract: "Abstract fluid art, iridescent colors, marble texture, gold accents, luxury aesthetic, macro photography --ar 1:1 --chaos 20"

4. Product Design: "Futuristic [product], minimalist design, white background, studio lighting, hyper-realistic, Octane render --ar 4:5"

5. Architecture: "Modern sustainable architecture, parametric design, green roof, golden hour, architectural photography, flowing parametric architectural forms --ar 16:9"

Include visual references you are allowed to use, clear constraints, and technical parameters that fit the image model.

When This Prompt Is Most Useful

Use this prompt when you want more control over AI image direction than a short aesthetic phrase can provide. It helps you define subject, composition, lighting, color, mood, aspect ratio, and exclusions before generating images.

The page is most useful for creators who need repeatable visual direction: product concepts, editorial images, mood boards, character exploration, environment studies, and campaign visuals. It is less useful when the goal is to imitate a protected brand or a living artist.

Inputs to Prepare

• Subject and purpose of the image
• Intended use: hero image, concept art, social post, ad, presentation, or reference
• Composition: close-up, wide shot, centered, rule of thirds, overhead, etc.
• Lighting and mood
• Color palette
• What to avoid: text artifacts, extra fingers, brand marks, clutter, unsafe elements
• Aspect ratio and model parameters

Example Input

Use case: landing page hero for a productivity app
Subject: focused freelance designer planning a project
Mood: calm, precise, warm
Composition: desk-level view, subject slightly left, clean negative space for headline
Avoid: neon AI clichés, fake app UI text, crowded desk, brand logos
Aspect ratio: 16:9

Useful Variations

• Art direction mode: Turn a rough idea into a visual brief with composition and exclusions.
• Parameter mode: Ask for aspect ratio, stylization, chaos, and quality suggestions based on the use case.
• Critique mode: Paste a generated result and ask what to change in the next prompt.

How to Review the Output

A useful image prompt should:

• define the subject and visual hierarchy clearly
• include exclusions, not only desired elements
• avoid copying named artists, protected brands, or private likenesses
• fit the final placement where the image will be used
• give you a way to iterate after the first generation

Follow-Up Prompts

• “Create three prompt variants: minimal, editorial, and cinematic.”
• “Add negative prompts to reduce text artifacts, extra limbs, and clutter.”
• “Rewrite this for a product hero image with space for headline text.”

Example Creative Workflow

Start with a plain brief before writing the image prompt. Define the subject, placement, audience, mood, and what should be avoided. Then generate three prompt directions: one minimal, one editorial, and one more cinematic. Compare the results against the actual use case instead of choosing the most dramatic image.

For example, a product hero image usually needs negative space, readable composition, and visual restraint. A concept art image can carry more detail and atmosphere. The same subject needs different prompt language depending on where the image will appear.

When Not to Use It

Do not use this prompt to imitate a specific living artist, recreate a protected brand identity, or generate a private person’s likeness without permission. Use broad visual language and references you are allowed to use.