Abstract: In the process of broaching, the natural frequency of the broaching machine greatly affects the dynamic characteristics of the machine, the machining quality of the workpiece, and the service life of the cutting tool. To ensure that the machine has good dynamic characteristics and machining accuracy while achieving lightweight design of the broaching machine, a multi-objective optimization design method based on response surface methodology is proposed. Firstly, parameterize the column of vertical broaching machine in SolidWorks and perform finite element analysis on the model using HyperMesh and Abaqus for pre-processing and post-processing. Secondly, the experimental design algorithm is used to design the combination of parameter variables, and the model parameters are automatically modified and finite element analysis is performed through script and batch processing methods to obtain the model quality, natural frequency, and maximum modal displacement corresponding to different parameter variable values for establishing the response surface. Finally, extract the response to establish a response surface model, and use the global response search method to solve under design conditions for multi-objective optimization. The research results show that after optimization, the quality of the column of machine decreased by 14.36%, the first-order natural frequency increased by 19.81%, the maximum displacement of the first-order mode decreased by 1.02%, and the maximum static deformation decreased by 2.33%.