Abstract: Aiming at the problems of abstract theoretical content of machine vision course, difficult to understand and lack of relevant experimental teaching devices, a license plate model positioning and character recognition experiment platform was designed based on machine vision, which was composed of lower computer control system and upper computer software. The lower computer is composed of image acquisition and transmission system and motion control system. The upper computer software is composed of human-computer interaction interface, positioning algorithm and character recognition algorithm. Based on VS2019 and Halcon, the control software is designed including target positioning and tracking, MLP classifier training, character segmentation and recognition and other functional modules. The location tracking and character recognition of license plate model are realized. The experimental design is practical, interesting and extensible, which can help students learn the real application scenarios of machine vision technology and improve their practical ability and innovation ability.