Aimed at bearing fault signal noise reduction processing, this paper studies the particle filter method and its application in signal noise reduction. Firstly, the mathematical model for bearing fault vibration signals is established and it is considered as the state equation of the particle filter. Then, extracting the background noise, the noise signal with the state signal as an observation signal, an observation equation is obtained. The original real signal is estimated, and the obtained estimate signal is the denoising signal. Through the simulation analysis, the signal-to-noise ratio is improved before and after noise reduction. Finally, the particle filter for the noise reduction is used for bearing fault vibration signal noise reduction processing in laboratory transmission boxes, thus obtaining same good effect of the noise reduction.