This paper studies the performance of wavelet transform denoising, and the algorithm of Cyclic autocorrelation analysis. Based on theoretical analysis, it makes Cyclic Autocorrelation function analysis of the measured multiple signals based on wavelet threshold denoising. From the results of simulation, it is effective in extracting bearing fault characteristic frequency and suppress interference from other frequency components by the cyclic autocorrelation function amplitude spectrum. The simulation results demonstrate the effectiveness of the method.