An improved active disturbance rejection controller (ADRC) was designed to solve the problems of slow response, low tracking accuracy and poor anti-disturbance ability of the electro-hydraulic position servo system of a mine sweeper launcher. BP neural network, which has powerful self-learning and nonlinear approximation ability, is used to adjust the key parameters of ADRC online, and combined with genetic algorithm (GA).The initial weights of the network are optimized, and AMEsim and Simulink software are used to co-simulate and verify the improved ADRC. The results show that the control method can effectively improve the anti-interference ability of the system, and ensure the speed and accuracy of gun adjustment of the mine sweeper.