Skip to content

Newbies, Don’t Miss Out! Master the Basic Structure of Midjourney Universal Prompts

A typical Midjourney prompt can follow this structure:

Subject : The main object or character you want to appear in the image.

Details & Setting: Describe the subject’s characteristics, emotions, actions, and the environment it is in.

Style: Specify the artistic style of the image, such as realism, cartoon, cyberpunk, etc.

Medium & Techniques: Although not always necessary, you can mention the painting style or technique you want the image to mimic, such as oil painting, watercolor, sketch, etc.

Parameters: Include specific requirements for the image, such as size, aspect ratio, and tone.

Writing Tips

  • Be Concise and Clear: Avoid long and complex sentences. Use short, powerful phrases.
  • Use English: Since Midjourney understands English better, write your prompts in English.
  • Avoid Prepositional Phrases: Try to use adjective + noun structures instead of prepositional phrases.
  • Weight Allocation: Place important descriptions at the beginning, as AI gives higher weight to words that appear earlier.
  • Draw Inspiration: Refer to successful prompt examples from others and adjust them with your own creativity.

Universal Prompt Structure

The order of keywords represents their weight, so the subject is always placed first.

Example: (A handsome guy) + (with black square hair, wearing suspenders) + (playing basketball on the basketball court) + (modern urban style) + (warm tones) + (cinematic lighting) + (--v6 --ar4:3)

Official Example

Green Part: Image link (for reference image, can be empty)

Blue Part: Descriptive keywords

Orange Part: Parameter settings

Example with a Pink-Haired Girl (in English)

(No reference image needed, leave this part empty) A girl with pink over-the-shoulder long hair and brown eyes, in the hotel banquet hall, modern urban style, warm colors, and movie lighting --v 6.0 --ar 4:3 --s 250

Note: The purpose of a reference image is to extract features such as style, color, and composition from an image to generate a similar one.