This system takes the warehouse monitoring system with intelligent tracking function as the construction goal, and conducts in-depth research on three neural network algorithms: CNN, Faster R-CNN, and Mask R-CNN to meet the needs of different industrial scenarios. The algorithm is deeply integrated with industrial application scenarios, and an artificial intelligence algorithm model for this system is established by combining computer vision technology. A solution for integrating artificial intelligence into an electronic fence system is proposed. The artificial intelligence algorithm model established in this article is used to perform deep learning and algorithm processing on the video captured by camera monitoring, replacing manual monitoring of warehouse items for abnormal movement. When warehouse items are moved, an intelligent alarm will be issued. This system solves the pain points of personnel on duty and the inability to intelligently identify various abnormal scenarios in warehouse monitoring, and realizes a new generation of electronic fence monitoring system with artificial intelligence characteristics.