Abstract: For the problem that the detection process of inner surface damage of tunnel based on acoustic reflection mechanism is susceptible to the interference of water environment noise, multipath noise and the noise of detection equipment, a denoising method of acoustic point cloud on the inner surface of water conveyance tunnel based on acoustic echo characteristics is proposed in this paper. The method utilizes the multi-echo data points formed by the inherent width characteristics of acoustic waves and the high-intensity values of data points formed by the superposition characteristics of acoustic waves, and combines the spatial position information of the point cloud to realize the noise filtering of the acoustic point cloud on the inner surface of the tunnel. Moreover, a series of experiments are conducted to verify the influence of different sensitive parameters on the denoising effect of the proposed method, and the results are compared with other classical point cloud filtering algorithms. The experimental results show that the proposed method is superior to the traditional laser point cloud filtering algorithm in the smoothing, denoising and reconstruction accuracy of the acoustic point cloud model on the inner surface of the tunnel. This method has practical research value for the detection of inner surface damage of water conveyance tunnel and the prevention of disasters in major water resources allocation projects.