Abstract: Using experimental dynamic response data as a reference, the inverse optimization of material parameters in collision simulation based on the finite element method and inverse optimization techniques provides an effective approach for obtaining high-precision base material parameters. However, traditional direct inverse methods based on simulation and optimization algorithms often suffer from low efficiency. To improve the efficiency of inverse optimization, this paper proposes a hybrid method that integrates Latin hypercube design, a radial basis function (RBF) surrogate model, and the NSGA-II algorithm. Taking compression test data of an energy-absorbing box as the reference, the accuracy and efficiency of the two inverse optimization methods were compared. The results show that both methods can meet the accuracy requirements for engineering applications. Compared with the direct method, the hybrid method requires fewer simulation model samples and achieves higher efficiency. The findings of this study provide a useful reference for the rapid and accurate acquisition of material parameters in collision simulation.