SOHL Multi-Dish YOLO is a food recognition dataset for multi-object detection tasks in computer vision. It is built based on the YOLOv8 annotation specification and focuses on the problem of detecting multiple dishes in complex scenarios. It aims to improve the model's object detection and classification capabilities in real catering environments. This dataset contains 377 labeled images, each with 377 corresponding annotations, covering 16 food categories. Each image contains 2–6 food objects, exhibiting characteristics such as overlap, multi-scale, and complex layout. The data is organized in YOLO format and includes image files and corresponding bounding box annotation text.