Abstract: Precision is the goal of machining field. CNC turn-milling compound machine tool as high-end machining equipment, the dynamic and static characteristics of its bed have significant influence on the machining accuracy. Aiming at the problems of low efficiency and poor effect of current bed structure optimization methods, a multi-objective joint optimization scheme based on nonparametric response surface method is proposed in this paper. According to this scheme, the optimal sizes of the bed model are selected based on the finite element analysis results. The mass of the bed, the maximum deformation and the first-order natural frequency are taken as the joint optimization objectives. Through the establishment of parameter test, the sensitivity analysis of sizes is carried out and the size with greater sensitivity is accurately selected. Based on the nonparametric regression method, the response surface is established. The selected sizes are imported and the response values of the test samples are calculated. The nonparametric regression response surface optimization model is established and the multi-objective genetic algorithm is used to find the optimal design scheme. The optimization result is as follows: the maximum deformation of machine tool bed is reduced by about 14.667%, the first-order natural frequency is increased by about 3.187%, and the quality is decreased by about 1.208%. The results show that the optimization scheme can effectively improve the dynamic and static characteristics of machine tool bed. This paper provides a theoretical basis for the optimization design of machine tool.