The operation and maintenance inspection of traditional power equipment has the characteristics of high repeatability, low staff efficiency and of being error-prone. An operation and maintenance system of secondary equipment based on Augmented Reality (AR) technology is constructed. It includes the standardized operation guidance database and the operation inspection resource database. A-KAZE feature extraction algorithm is introduced and clustering analysis and point set verification are combined to give full play to the technical characteristics of the operation and maintenance system of secondary equipment in image recognition, virtual and real image overlay, intelligent error correction judgment and gesture voice interaction via the implementation of OpenCV source code. After capturing the information of the inspected equipment, the AR intelligent terminal equipment can automatically detect abnormal equipment by comparing the call with the background information database. It can also record the on-site inspection data in real time. The application results in the operation and inspection of secondary equipment in a substation show that, the constructed system standardizes the operation process of transportation inspection, realizes paperless operation, and improves the inspection efficiency and identification accuracy of secondary equipment in a substation. This work is supported by National Natural Science Foundation of China (No. 51477061).