In order to reduce the interference of noise on aeroengine mechanical fault diagnosis result and improve the classification performance of fault feature set, an integrated fault feature set extraction method based on stochastic resonance(SR) was proposed. First, the stochastic resonance was applied to pretreat the vibration signal, thus improving the Signal to Noise Ratio(SNR) and enhancing the frequency characteristics of the output. Then, the fault feature set was extracted from the output signals of SR system. The fault feature sets based on time domain analysis, frequency domain analysis and time-frequency domain analysis was proposed respectively to test the treatment effect of SR method. And the rotor test data was used to test the extracted feature sets. The results indicate that the fault feature set extracted from the output signals of SR system showed the better classification performance and the diagnosis result had the higher stability than the feature set extracted from the original signals.