Abstract: A multi-objective optimization design method combining numerical simulation and optimization algorithms is proposed to address the problem of difficulty in handling the correlation between various design parameters in the optimization process of complex machine tool structures. The SolidWorks and Workbench plug-ins are called in the Isight platform for experimental design. The sample points are designed by the optimal latin hypercube algorithm (OLHA), and the universal Kriging approximation model is constructed with the generated sample points. The reliability and effectiveness of the model are proved by cross validation. With the minimum maximum displacement deformation and the maximum first-order natural frequency as the optimization objectives, the multi-island genetic algorithm (MIGA) is used for optimization calculation to obtain the optimal solution. The optimized first-order natural frequency is 221.31 Hz, an increase of 8.43% compared to before optimization, and the maximum deformation is 0.001888 mm, a decrease of 18.45%, achieving multi-objective optimization design of the rack.