Aiming at the problems of low path tracking accuracy and external uncertain interference in the cleaning process of photovoltaic array cleaning robot, an improved active disturbance rejection control strategy is proposed to control the driving unit model, which realizes the high robustness control of the angular velocity (torque) of the driving unit and improves the path tracking accuracy of the robot. Error between actual and expected motion poses of a cleaning robot by analyzing the motion state of the robot. Due to the interference of the external environment and other uncertain factors, by establishing the dynamic control model of the mobile chassis of the cleaning robot with uncertain interference factors, and by improving the fal function on the basis of the traditional ADRC, an improved ADRC strategy with embedded kinematics and dynamics is proposed. Cleaning robot realizes high precision tracking target path by improved extended state observer observing uncertain disturbance factors in real time. The robot showed good tracking results through a variety of target path tracking simulation experiments. Therefore, it proves the superiority and effectiveness of the path tracking control algorithm of photovoltaic array cleaning robot based on improved auto disturbance rejection control, and improves the path tracking accuracy of photovoltaic array cleaning robot.