Abstract: For the detection of small low speed targets with weak echo intensity, such as frogman, under the background of random noise and reverberation in shallow sea, a target detection method based on sonar history accumulative image is proposed in this paper. First, according to the correlation of sonar images in time domain and spatial domain, the space-time normalization processing technique is used to suppress strong interference to the echo such as static reverberation and bursting noise in the sonar background. The sonar history accumulative image integrates the information of multiple frames of sonar images, and the echo highlights of target form bright line features due to moving continuity. By this feature, Radon constant false alarm rate (Radon-CFAR) method is used to detect the short-term moving trajectory of the target in the sonar history accumulative image, and the target with low signal-to-noise ratio can be detected. The performance of the space-time normalization processing technique and the detection algorithm are analyzed, and the feasibility of the detection algorithm is verified with sea trial data, which shows that the weak echo of low SNR from frogman target can be well detected.