As the proportion of photovoltaic power generation continues to increase, the number of photovoltaic panels installed is also increasing. In response to the direct impact of dust and sand blocking the surface of photovoltaic panels on power generation efficiency, as well as the high cost and low efficiency of daily manual maintenance, this paper designs a photovoltaic panel surface cleaning robot and elaborates on the main structure of the robot: cleaning module and moving module. In order to enable the robot to efficiently complete cleaning work, the robot’s movement trajectory on the photovoltaic surface is planned. The planning method adopts a control strategy combining neural network and PID. Then, the machine’s control system is designed based on the planning method. Finally, through experimental verification, the rationality of the robot’s cleaning structure design is analyzed and verified, and the cleaning efficiency of the robot is improved.