Abstract: Compared with traditional single-decker ball bearing，the double-decker ball bearing has higher maximum speed and longer service life. However，the use of double-decker ball bearing is constrained due to its large mass. To solve the problem，the structure of double-decker ball bearing is optimized. The equivalent stress，contact stress，and temperature rise of the bearing are obtained through finite element analysis. The local sensitivity of each design variable is analyzed based on the response surface method，and the parameters that are highly correlated with the objective function are determined.The multi-objective optimization design of double-decker ball bearing is carried out using the multi-objective genetic algorithm，in which the bearings' mass，fatigue life and temperature rise are takren as the optimization goals，and the ball diameter，number of balls and coefficient of groove curvature radius are taken as the design variables. The optimal internal geometric parameters of double-decker ball bearings were obtained. A test device was built according to the optimized results.The results show that the mass of the optimized double-decker ball bearing is reduced by 10.47%，the maximum contact and equivalent stresses are decreased by 9.11% and 6.45%，repectively，and the maximum temperature is increased by 1.9%.

Key words: double-deckerballbearing, responsesurfacemethodology, multi-objectiveoptimization, staticanalysis, steady-statethermalanalysis