Aiming at the problem that the rotor position eccentricity error was increased after the rotor of permanent magnet AC torque motor was loaded,a method based on genetic algorithm (GA) and BP neural network was proposed to optimize the position error prediction model of rotor under eccentric load.By measuring the position change of the permanent magnet torque motor in the running process under various conditions,the genetic algorithm was used to optimize the BP neural network algorithm to establish the prediction model.In the model,the sinusoidal trajectory data run by the experimental bench were used as the training sample,and the triangular wave trajectory data were used as the test sample.The sine wave trajectory data and the triangular wave trajectory data of various situations were selected for simulation prediction and verification.The command position,command speed,and current of the sine wave signal in various cases were used as the input of the model,and the position error of the triangular wave signal was used as the output.The results show that the prediction model established by GA optimization BP neural network is better than the prediction model before optimization.