"Intelligent garbage processing platform based on AI vision model and dynamic planning algorithm" uses mature intelligent technologies such as artificial intelligence, big data and IoT to create new products and services in the whole garbage classification chain. The system includes an intelligent sanitation transportation management system that integrates the web page management end and the small program end of garbage removal personnel, and an embedded device that deploys the intelligent visual model Swin Transformer, which realizes the intelligent co-ordination and supervision of the whole process of garbage removal, and adopts a dynamic planning algorithm, which greatly improves the execution efficiency of garbage removal. At present, the system has obtained one soft work and applied for one national invention patent. The main highlights of the system are as follows:(1) the task is highly streamlined, the administrator can directly view the task progress and obtain relevant information on the web page, and at the same time split the task into multiple nodes for dynamic distribution, thus greatly improving the work efficiency and flexibility;(2) Introduce the concept of default tasks and refresh them regularly every day to fully improve the management efficiency;(3) View the location of the task in real time to ensure the smooth progress, fully improve the collaborative ability;(4) Investigate the mainstream in-depth learning vision model in academic circles and make appropriate improvements on the basis of full research. The accuracy rate of garbage classification and identification can reach 90%, thus realizing the accuracy of garbage classification. (5) The model is deployed to strawberry Pi IoT devices to realize intelligent acquisition of dynamic images, enrich the training library and continuously improve the accuracy of model identification;(6) Use dynamic planning algorithm to efficiently calculate the optimal path of multi-task in real time, save detour time, greatly improve the efficiency of task execution.