In view of an improvement of the service performance of deep groove ball bearings, an optimization design method has thus been proposed based on particle swarm-genetic hybrid algorithm. With rated dynamic load and rated static load as the objective function, and with the diameter of rolling elements, pitch circle diameter, number of rolling elements, and curvature radius coefficient of inner and outer raceways as the design variables, based on the particle swarm optimization, penalty functions and genetic crossover and mutation operations are introduced for the solution of constrained optimization problems and local optimization problems. Taking 6206 bearing as a calculation example, a stress and sensitivity analysis is carried out for the optimized bearing. The results show that the proposed algorithm is characterized with an improved convergence performance, a stronger optimization ability, and a faster computational speed. The optimized deep groove ball bearing contact stress has decreased by 31.7%, thus verifying the validity of the proposed method.