In order to better apply the real-time control algorithm to the automatic navigation of unmanned ships, a remote control system based on user datagram protocol(UDP) is established according to the Japanese national research institute of fisheries engineering (NRIFE), the system uses python on the server side and Matalb/Simulink on the client side. Aiming at the problems of inconsistent data types and unreliable communication of cross-platform real-time control in the past, specific solutions are proposed. In order to verify the real-time performance of the control, a real-time course control experiment based on the active disturbance rejection control (ADRC) technology is carried out in the experimental tank, which verifies the real-time performance of the designed control system and the effectiveness of the control algorithm. When testing the system, the simulation of the unmanned ship dynamic positioning system is carried out using the RBF neural network disturbance rejection algorithm, and the stability of the controller is analyzed. The experimental results show that the UDP communication of remote control is smooth and the control results are satisfactory, which provides the foundation for subsequent complex unmanned ship experiments.