For CNC machining, surface texture of a workpiece can be generally divided into overcutting, chattering, oblique slope vibration marks, cutting ring, wiring, etc. Analyzing and interpretation of the surface texture usually require an experienced engineer, and the interpretation result for workpiece under different lighting conditions and different angles will also be different. This article described how to combine AI (Faster RCNN) and image technology to classify the surface texture type automatically. Users only need to put the processed standard mould part in the designed camera module then the texture type will be automatically marked. In the future, experts will be combined with Internet of things (IoT) systems to achieve the goal of automatic determination of the cause of surface texture.