In order to improve the heat dissipation performance and energy density of power lithium battery packs, based on the semiconductor refrigeration scheme, a multi-objective optimization design method is proposed to optimize the arrangement of power lithium battery packs and semiconductor cooling capacity. Based on the established thermal analysis model of the semiconductor refrigeration scheme, the Latin hypercube test, radial basis function (RBF), response surface methodology (RSM), and Kriging proxy model methods are used to establish the approximate model of the maximum temperatures, maximum temperature difference and pacing volume. Taking the maximum temperature difference and spacing volume as the goal and the maximum temperature as the constraint, the battery pack heat dissipation optimization model is established. Then, the optimization solution is found by using the multi-objective genetic algorithm. Finally, and the reliability of the simulation results of the optimization scheme is verified through experiments. The simulation results after optimization show that the battery module spacing volume is reduced by 32.42%, the maximum temperature difference is reduced by 13.64%, and the maximum temperature is reduced by 0.53%. This method significantly improves the heat dissipation performance and energy density of the battery packs.