In this paper, the functional design of intelligent cleaning robot is studied in depth. Ankai micro development board is used as the core, and advanced technologies such as machine vision and laser radar are integrated to realize accurate garbage identification and classification. Through the lightweight YOLOv5s neural network, the rapid detection of objects is realized, and the mechanical arm is used to perform garbage picking. The random coverage method is used for path planning to ensure comprehensive cleaning, and the improved Bug2 algorithm improves the efficiency of obstacle avoidance. The system design aims to improve the cleaning efficiency, reduce the cost and promote environmental protection, which indicates the wide application prospect of intelligent cleaning robot in the future.