Midjourney V5 prompt work improves when the brief makes the subject,
composition, visual language, and constraints explicit. Treat model
names and parameters as version-sensitive: verify which settings are
available in your current account before relying on an older
tutorial.
A reusable structure
[subject] in [setting], [composition and camera or viewpoint],
[medium or visual language], [lighting and palette], [mood],
[important details], [things to avoid], [verified parameters]Ten prompt formulas
- Editorial portrait:
[person or character], [expression], [wardrobe], editorial portrait, [lens/viewpoint], controlled studio light, [palette], clean background - Product concept:
[object] on [surface], [materials], three-quarter view, soft directional light, realistic proportions, uncluttered product-study background - Environment:
[place] at [time], wide establishing view, foreground/midground/background separation, atmospheric perspective, [mood], no readable text - Architectural study:
[building type], [architectural features], human-scale reference, orthographic-inspired composition, [materials], neutral daylight - Cinematic frame:
[subject] during [action], cinematic composition, [shot size], motivated lighting, controlled color palette, clear focal point - Botanical illustration:
[plant], identifiable anatomy, [illustration medium], restrained palette, clean negative space, no lettering - Character sheet:
[character], front/side/back reference views, consistent clothing and proportions, neutral background, no extra characters - Texture study:
[material], close-up surface study, raking light, macro detail, even focus, no objects or logos - Poster concept:
[event or theme], strong central symbol, balanced margins, limited palette, leave space for later typography, no generated words - Variation set:
[subject] shown in three clearly different compositions, consistent identity and palette, labeled only by position, no readable text
Improve one variable at a
time
Keep the subject and composition fixed while testing lighting,
medium, palette, or camera angle. Save the best prompt and the change
that improved it. This turns image generation into a repeatable
experiment instead of random prompt swapping.
Review anatomy, text artifacts, trademarks, recognizable styles, and
commercial rights before using an output. For more reusable workflows,
explore the ChatGPT
blogging prompts and Complete
Playbook Bundle.