Distributed optical fiber acoustic sensing (DAS) is a kind of optical fiber sensing system based on phase sensitive optical time domain reflection technology (Ф-OTDR). It has the characteristics of high sensitivity, long distance and real-time detection, and has been widely used in the positioning and monitoring of various engineering events. However, the high coherent detection light source used in the Ф-OTDR technology will produce coherent fading phenomenon in the fiber, which causes fading noise. This will seriously affecting the monitoring effect of the DAS system. Therefore, in order to address the fading noise problem in DAS system, a system fading noise suppression method based on non-local means filtering is proposed, which can effectively suppress the system fading noise in the signals and improve signal-to-noise ratio while preserving the characteristics of the target event signals. The experimental results show that the proposed method has better DAS system noise suppression effect in the face of different intensities and types of noise interference. Further, the proposed method is compared with EEMD, wavelet and outlier noise suppression methods. The results show that the noise suppression effect is significantly better than other methods, and the signal-to-noise ratio is improved by more than 30%, which verifies the effectiveness and superiority of the proposed method.