A baggage detection system based on deep learning is designed to solve the problem of high security pressure faced by large places. The core of the system adopts VGG-16 and SSD framework. According to the requirements of security inspection, it can track, locate and identify the baggage that has passed the X-ray scan of security inspection machine. The results are presented to users through the alarm mechanism of the system, and real-time statistics and evaluation are carried out. The system is divided into three modules: detection module, Java background module and Web foreground module. The practice results show that the system has a better identification effect and can meet the production requirements on site.