Abstract: In view of the defects of the cylinder wall in the actual production of cold extrusion of aluminum alloy accumulator shell parts, multi-objective optimization was performed based on the combination of finite element software DEFORM-3D and response surface method and multi-objective optimization NSGA-II analysis. Firstly, the AA6061 aluminum alloy bar was subjected to room temperature tensile experiment to obtain the stress and strain data, and then imported into DEFORM-3D to construct the FEM model. Secondly, with the transition fillet X₁, extrusion speed X₂, and friction coefficient X₃ of the punch working part as optimized variables, a mathematical model of extrusion load Y₁ and surface damage Y₂ of the shell parts was established. The variance results showed that the model had high accuracy and was very effective. A good description of the response of the two optimization goals to the design variables, and the 3D response surface diagram could visually analyze the conflicts between the extrusion load and the surface damage of the part with respect to the response variable. To resolve these, NAGA-II was used for multi-objective optimization, and a set of Pareto optimal solutions was obtained. Then, a reasonable range of process parameters for extrusion forming was obtained: X₁ was 0.64~0.68 mm, X₂ 5.8~6.2 mm/s, and X₃ 0.1. Finally, a set of optimal parameter combinations was selected for experimental verification, and the results showed that the work-piece forming performance and quality were good, and the simulation results and experimental results had good reliability.