A set of UVW positioning system based on machine vision was developed to meet the requirement to high performance platform.A method of visual system calibration based on machine learning was presented.Based on the platform coordinates and the coordinates of the image obtained from the kinematic analysis of the platform,the inverse kinematics of the UVW platform was established.The machine learning visual calibration and binocular vision positioning control experiments were accomplished.The results show that the designed visual calibration method of UVW platform is simple and efficient,and the motion platform has high accuracy,which can meet the needs of industrial applications.