Abstract: The transmission efficiency of the reducer is an important factor to improve the efficiency of the drive axle transmission system of electric vehicles. Traditional efficiency modeling is difficult to reflect the actual situation of the transmission system under complex working conditions, so the GA-BP neural network is used to establish the transmission efficiency model of electric vehicle reducer. Compared with the traditional efficiency model, the model can comprehensively find the optimal solution through genetic algorithm, and use the nonlinear ability of neural network to construct the complex relationship between lubricating oil temperature, input torque and speed and the transmission efficiency of the reducer. The simulation results show that the RMSE of the model is 0.24354 and 0.33229, respectively, and the maximum relative error percentage is less than 4%, which verifies the accuracy of the model. Therefore, the model can well reveal the variation law of the transmission efficiency of the reducer with working conditions.