The problem of cable insertion for different types of distribution terminal FTU (Feeder Terminal Unit) in the detection is increasingly prominent. An automatic recognition system of socket status based on machine vision was designed to solve the problem above. The system consists of an industrial computer, an image acquisition processing module and an industrial robot. The image is collected and processed by CCD camera and image acquisition card. The details of image processing techniques such as median filtering, grayscale, distortion correction and camera calibration are utilized to simplify the following calculation. The calculation method of position and angle deviation and the method of vision characteristics of FTU navigation socket are proposed. The system solves the problem of poor accuracy and low efficiency of insertion in the automatic inspection process of FTU, improves the automation degree, and avoids the safety hazard of manual live working.