In order to study the effects of bounded-noise parameters’ change on system stochastic resonance（SR）， the outputs of three typical systems are investigated. Stochastic Runge-Kutta method is used to discretize the equations and calculate system response in a extend period. The results demonstrate that the SR can be effectively improved by changing the mean frequency and parameters of the bounded noise， and the resonance effect can be optimized by the existence of numerical relations. With the increase of mean frequency， the peak value of SR will increase and deviate， but the enlargement of will suppress the phenomenon of SR. To verify the above results， we establish Simulink models， and the same conclusion can be obtained through the simulation experiment.