For the purpose of optimizing the parameters of electric locomotive traction electric drive system for more reasonable matching relationship, an automatic optimization method for the key parameters of electric locomotive traction electric drive system is studied. Firstly, the traction electric drive system simulation model of a certain electric locomotive is established, and the simulation results prove the accuracy and rationality of the modeling method. Secondly, simulation based on design of experiment (DOE) of the model is carried out, and a cluster of simulation results reflecting the system operation law is obtained. Based on the simulation results, an approximate model mapping the input parameters and output characteristics of the system is established. Thirdly, the optimized design parameters are obtained by the multi-objective particle swarm optimization algorithm. Finally, the results of simulation optimization of several models are compared and analyzed, which verifies the effect of the global optimization method based on the approximate model, and high simulation efficiency is achieved as well as high simulation accuracy. The research has certain reference value for the high-efficiency simulation of electric locomotive traction electric drive system.