Based on deep learning, front-end and back-end technologies, the work realizes insect recognition of user-end and management-end modules, query and addition functions of insect database, and is equipped with a good-looking UI interface and improves the deep learning recognition algorithm. The user side realizes the function of taking photos and identifying pests, and the user can also carry out fuzzy query on pests for the insects he needs. In addition, we provide a worm house module, where users can contribute, view and collect articles on pest control or plant cultivation experience, and can use the suggestion box for feedback. The management side has small programs and PC ends. Small terminal: administrators can easily add and modify the data of pests. View the content of comments submitted by users. In addition to the basic functions of the small program terminal, the PC terminal can also manage the menu bar in the worm house and review the articles submitted by users.