Abstract: Refrigerator-cooled cryopumps have been widely applied in fields such as nuclear fusion, space environment simulation, and high-energy physics research owing to their advantages, including oil-free cleanliness and high pumping speed. To meet the demand for improving the pumping performance of cryopumps, a multi-factor and multi-objective optimization method based on orthogonal experiments and the entropy weight method is proposed. This method is employed to investigate the influence laws of parameters such as baffle angle ( \varphi ), baffle overlap ratio ( \eta ), number of cold plate layers (n), and cold plate radius (r) on the performance indicators of refrigerator cooled cryopumps, and to conduct a multi-objective comprehensive evaluation of cryopump performance. The results show that different parameters exert varying degrees of influence on each performance indicator. Furthermore, the weights of the above parameters relative to the pumping performance indicators are determined via the entropy weight method: secondary radiation heat has the largest weight, followed by primary radiation heat, the maximum temperature of the cold plate array, and transmission probability. On the basis of these weights, the structure of the cryopump is optimized using the comprehensive evaluation and analysis method of orthogonal experiment, and the optimal parameter combination scheme for the cryopump is obtained: baffle angle of 50°, baffle overlap ratio of 0.9, cold plate layers of 9, and cold plate radius of 90 mm. The aforementioned multi-objective optimization method and comprehensive evaluation idea provide a useful reference for the pumping performance analysis and structural optimization of refrigerator-cooled cryopumps.