Abstract: With the deepening of smart grid construction, the demand for real-time, accuracy and full coverage of substation inspection is becoming increasingly urgent. This paper proposes a substation joint inspection data processing platform based on edge computing, which integrates the vibration, temperature, image and other multi-source data collected by multiple types of intelligent devices in the perception layer, and deploys typical algorithms such as signal processing, risk assessment, image recognition and data preprocessing at the edge computing layer to realize localized real-time analysis and decision-making of data, and completes global data management and in-depth optimization through the cloud layer. The system test shows that the data processing delay of the platform is controlled within 3s, the average accuracy of defect identification reaches 93.2%, and the inspection efficiency is increased by about 65%, which significantly reduces the operation risk of the substation and has good engineering application value.