In order to improve the performance and reduce mass of an ossicular silencer, the average transmission losses (TL) in the low frequency range and the broadband were chosen as optimization goals, several structural parameters (width of the main duct，depth of the enclosed cavity, and length of the enclosed cavity) were chosen as optimization variables, and the mass of the silencer was taken as the constrain condition. The finite element model was also established for the silencer, sample points were obtained by central composite design (CCD)，the quadratic polynomial model was constructed based on response surface methodology (RSM)，and the fitting accuracy was tested. Response surface methodology combined with genetic algorithm II (NSGA-II) was applied to optimize the structural parameters of the silencer. The optimized ossicular silencer shows better performance and lighter weight. The optimization results indicate that the structural parameters of the silencer are optimized by combining the response surface methodology with genetic algorithm.