# 搞笑AI图片提示词：用自拍和宠物照片做的9个点子，附结果

> 搞笑AI图片提示词靠的是同一个办法：把荒唐的场面拍得一本正经。我们用虚构人物Tim的自拍和他的柯基照片测了九个点子——从文艺复兴肖像到骑着狗的骑士——另外还有三组不靠原图的动物场面。下面是提示词、结果，以及怎样把图片做成搞笑视频。

发布于 2026-10-30 · 更新于 2026-10-30
URL: https://ai2face.com/zh/blog/funny-ai-prompts

搞笑的AI图片不是靠在提示词里写上“funny”这个词。模型听不懂笑话，它听得懂描述。荒唐的场面写清楚，再拍得一本正经，才会好笑：戴着拉夫领的贵族抱着狗、骑着柯基的骑士、粉色浴袍里的超级英雄。我们用虚构人物Tim和他的柯基的照片测了九个点子。每张图都在 AI2Face 生成器里用 [Nano Banana Pro](/zh/models/nano-banana-pro) 完成。

## 怎么用搞笑AI图片提示词做一张照片

1. **选一张原图。** 白天的正面自拍；宠物则用正面全身照。
2. **选一个能按照片改图的模型。** 它会把脸——或者口鼻——带进新场面。
3. **贴上提示词**，结尾加上英文这句“keep his face exactly as in the photo”，再写上辨认特征：胡子、眼睛颜色、眼镜。

![左边是穿灰T恤、留着红胡子的男人，右边是木地板上的红白柯基](https://ai2face.com/_astro/sources.IJ_P4ra__ZgdFGs.jpg "原图：Tim的自拍和他的柯基照片。两者都是虚构的，为这次测试生成")

## 搞笑提示词的秘诀：用正经的拍法

这份清单里几乎每个成功的点子都是同一个公式：**荒唐的场面 + 严肃的类型 + 人物自己完全当真**。提示词里写“epic”“solemn”“cinematic”“museum lighting”，比写“funny”或“hilarious”更好笑。你要一张“funny picture”，模型就给你搞怪的脸和卡通风格——看着无聊。

## 1. 文艺复兴贵族，狗也戴着拉夫领

![金框油画肖像：留胡子的男人戴着巨大的白色拉夫领、穿着丝绒上衣，抱着一只同样戴拉夫领的柯基](https://ai2face.com/_astro/renaissance.CAP62-Oi_maPNa.jpg "16世纪肖像：两者都极其严肃地看着观众。Nano Banana Pro")

```
A Renaissance oil portrait in an ornate gilded frame: the man from the
first photo as a pompous 16th-century nobleman in a huge white ruff
collar and a velvet doublet, solemnly holding the corgi from the second
photo who also wears a tiny ruff collar, dark background, museum
lighting, cracked varnish texture, both looking at the viewer with total
seriousness. Keep the man's face and the corgi exactly as in the photos.
```

## 2. 桌上的摇头娃娃

![收藏摆件：留胡子男人的脑袋特别大，身子很小，放在办公桌键盘旁边](https://ai2face.com/_astro/bobblehead.BNPHToOl_2shRyx.jpg "摇头娃娃：脑袋是身子的三倍，脸仍然认得出。Nano Banana Pro")

```
A collectible bobblehead figurine of the man from the photo with a
comically huge head and a tiny body in a grey t-shirt, standing on an
office desk next to a keyboard and a coffee mug, glossy painted vinyl,
shallow depth of field, product photo. The figurine's face clearly
resembles him.
```

## 3. 骑着柯基的骑士

![一身闪亮铠甲的骑士骑着同样穿着小铠甲的柯基冲向战场，背后是落日天空](https://ai2face.com/_astro/knight.B89xh3nx_ZzdXmv.jpg "史诗冲锋：正经的场面反而好笑。Nano Banana Pro")

```
Epic heroic scene: the man from the first photo as a medieval knight in
shining armor charging into battle riding the corgi from the second
photo as his noble steed, the corgi in tiny knight armor with a flowing
banner, dramatic sunset sky, dust flying, cinematic wide shot,
completely serious epic mood. Keep the man's face and the corgi as in
the photos.
```

## 4. 街头画家的漫画脸

![水彩漫画脸：巨大的红胡子、又大又圆的脑袋、小小的身子](https://ai2face.com/_astro/caricature.DFlz84K8_Z1yqjQV.jpg "漫画脸：模型自己挑了最显眼的特征——胡子——然后把它放大。Nano Banana Pro")

```
A funny hand-drawn caricature of the man from the photo: exaggerated
huge ginger beard, big round head, tiny body, cheerful grin, watercolor
and ink style like a street artist's caricature on white paper,
recognizable likeness. No text, no signature.
```

更多漫画脸和卡通样子，看我们的[AI动画生成](/zh/blog/create-cartoon-ai)指南。

## 5. 浴袍超级英雄

![男人把粉色浴袍当披风，穿着兔子拖鞋，头上扣着漏勺当头盔，爬梯子去救树上的猫，邻居在笑](https://ai2face.com/_astro/superhero.DHBzHUIz_ZK4Dli.jpg "救猫：浴袍当披风，漏勺当头盔。Nano Banana Pro")

```
The man from the photo as a superhero in a fluffy pink bathrobe used as
a cape, bunny slippers and a kitchen colander as a helmet, heroically
climbing a ladder to rescue a grumpy cat from a tree in a suburban
backyard, neighbors watching, sunny day, cinematic action photo.
Keep his face exactly as in the photo.
```

## 6. 巨大三明治上的迷你自己

![一个杯子大小、留着胡子的小人坐在巨大的三明治上，把面包屑举得像奖杯](https://ai2face.com/_astro/tiny-me.f6MB2ovS_Z12ALsj.jpg "迷你自己的微距：比例可信，脸也认得出。Nano Banana Pro")

```
A tiny miniature version of the man from the photo, the size of a coffee
cup, sitting cross-legged on top of a giant sandwich on a kitchen table,
holding a crumb like a trophy, macro photography, shallow depth of
field, realistic scale. Keep his face as in the photo.
```

## 7. 柯基首席执行官

让宠物扮演人，是最稳的办法之一。有一张狗或猫的照片就够了。

![柯基穿着西装打着领带坐在会议桌主位，两侧各有一只穿外套的金毛寻回犬](https://ai2face.com/_astro/corgi-boss.DCYBexBX_Z20xoUT.jpg "董事会：柯基对数字很失望，金毛们很紧张。Nano Banana Pro")

```
The corgi from the photo as a strict CEO sitting at the head of a long
boardroom table in a tiny navy suit and tie, paws on the table, looking
disappointed at a presentation chart, two golden retrievers in suits as
nervous employees, corporate office, photorealistic, dramatic office
lighting. Keep the corgi exactly as in the photo.
```

## 8. 仓鼠面试

连原图都不用——动物处在人的处境里，本身就好笑。

![仓鼠穿着西装，紧张地拿着简历，对面是戴眼镜、表情严厉的猫面试官](https://ai2face.com/_astro/hamster.CEevLhFM_QTYOg.jpg "面试：仓鼠脸上有一滴汗，猫面试官不为所动。Nano Banana Pro")

```
Photorealistic photo of a hamster in a tiny business suit at a job
interview, sitting on a chair across a desk from a stern cat interviewer
in glasses, the hamster nervously holding a tiny resume sheet with no
readable text, office setting, cinematic lighting, funny serious mood.
```

## 9. 健身鹅

![健身房里一只肌肉发达的鹅戴着头带，用嘴叼着杠铃，其他鹅在加油](https://ai2face.com/_astro/goose.CQtVNUah_2anuzM.jpg "鹅的励志现场：汗、头带、欢呼的观众。Nano Banana Pro")

```
Photorealistic photo of a muscular goose in a gym lifting a tiny barbell,
wearing a sweatband, other geese cheering in the background, gym
lighting, sweat drops, epic motivational mood.
```

模型自己决定让鹅用嘴叼杠铃——结果更好笑。动物场面很适合做成[AI做表情包贴纸](/zh/blog/stikery-nejroset)和[AI制作贺卡](/zh/blog/otkrytki-nejroset)。

## 把搞笑图片做成视频

这份清单里的任何一张图都能动起来：图片变成第一帧，提示词只写一个动作。柯基总裁用爪子拍桌子，鹅把杠铃掉下去，贵族从肖像里眨一下眼。4–8秒、只有一个动作的短片，比长片更干净。做法见[照片变视频](/zh/blog/ozhivit-foto)。

## 总结

搞笑AI图片提示词靠的是反差：类型和场面越正经，荒唐的处境越好笑。拍自己的脸，用能按照片保住相貌的模型，并加上“keep his face as in the photo”——我们的测试里，Tim在六个场面都认得出。宠物梗图有一张动物照片就够，办公室里的动物只需要文字。先从贵族和他的狗，或者骑柯基的骑士开始：这两个几乎谁看了都会笑。更多点子见我们的[AI绘画提示词](/zh/blog/promty-dlya-kartinok)和[AI 照片提示词模板](/zh/blog/gotovye-promty-dlya-foto)。

## 常见问题

### 能用我自己的自拍做搞笑AI照片吗？

能。上传自拍，选一个会按照片改图的模型，比如 Nano Banana Pro，再贴上以“keep his face exactly as in the photo”结尾的提示词。在我们的测试里，Tim在六个场面里都认得出，摇头娃娃也一样。

### 能免费做搞笑AI图片吗？

能做几张。AI2Face 注册后赠送欢迎积分，每次生成的费用在开始前显示。

### 怎样用一张照片做成搞笑视频？

把做好的搞笑图片当作第一帧，只描述一个动作：柯基总裁用爪子拍桌子，鹅把杠铃掉下去。4–8秒、只有一个动作的短片，比长片更干净。

### 做搞笑图片用哪个AI好？

用自拍做搞笑照片，选能按参考图保住脸的模型：Nano Banana Pro 能保住最多5个人的相貌。没有原图的动物梗图，强一点的图像模型都可以。

### 能用别人的脸做搞笑照片吗？

只有对方同意才行。没问过就拿朋友开玩笑，很容易伤人；拿名人的脸，还可能侵犯肖像权。

## 资料来源

- https://blog.google/innovation-and-ai/products/nano-banana-pro/
- https://ai.google.dev/gemini-api/docs/image-generation