On the basis of analyzing the noise characteristics of passive layer particles, a square-wave signal with low SNR was simulated. Based on the variation law, a Mallat fast algorithm was used to analyze the wavelet signals with low SNR and analyzed the characteristics of the noise. An algorithm designed to suppress interference noise in the passive layer particle temperatures. The simulation on the designed algorithm shows that the algorithm can filter the noise in the signal. By constructing a roller experiment device,the particle temperature of the passive layer of the roller is measured, and the measurement data is analyzed. The change of internal particle temperature state is effectively measured, and the practicality of denoising the passive layer particle temperature signal using wavelet transform is proved.