[Last edited: 5/25/2026]
<aside> <img src="/icons/magic-wand_gray.svg" alt="/icons/magic-wand_gray.svg" width="40px" />
There is no single “right” way to prompt in Midjourney. If the image works, the prompt works. BUT. When the image does not match your idea, prompt troubleshooting becomes easier if you treat the prompt as a visual control tool. Describe what should be visible. Put the most important subject early. Keep sentences short, direct, and easy to parse. Use concrete visual language instead of instructions, vague mood, camera metadata, or storytelling filler.
</aside>
Learn from this page or explore the GPT or GEM (for paid accounts only): Clarinet's Prompt Helper (ChatGPT) Clarinet's Prompt Helper (Gemini/Gem)
Midjourney v8.1 can handle more detail than earlier models, but it is also sensitive to wording, token order, punctuation, and ambiguity. If your prompt is too long or overloaded, Midjourney may shorten it, weaken important details, or invent missing information. Keep the prompt clear enough that the important parts survive.
A strong v8.1 prompt usually follows this shape:
Main subject with its most important visible traits. Closely attached subject details. Foreground or action. Background or environment. Style or medium. --ar 3:2 --v 8.1
Example:
A red fox with wet fur stands on a mossy riverbank. Its paws press into dark mud near scattered yellow leaves. A misty pine forest fills the background. Natural wildlife photography, soft overcast light. --ar 3:2 --raw --v 8.1
<aside> 💡
For more about v8.1, you can download the Midjourney Quick Start (v8.1)
For more about the changes from earlier versions, try V8.1 Community Release Notes
</aside>
If details are being dropped or the camera isn’t focusing on the right subject, you can explore changing the order of the details in the prompt.
For example, try putting the intended focus of the image earlier in the prompt, or first. Midjourney v8.1 gives earlier words (very slightly) more influence, so the first noun phrase could be the thing you most want the image to get right.
The grammatical subject of the sentence also changes the image’s focus. These two prompts contain the same basic ingredients, but they ask Midjourney to organize the image differently:
A duck standing in a field. --ar 3:2 --v 8.1
A field with a duck standing in it. --ar 3:2 --v 8.1