Underwater target detection, recognition and tracking are most important hot research questions, which have important application in military and civil fields. This paper aims at giving a comprehensive review of advances in the principles, methods and typical algorithms of sonar-image- based underwater object detection, recognition and tracking. We first make a thorough description of the main development, typical algorithms and their extensions of underwater object detection, sonar image denoising and segmentation. Then, the main methods and technical difficulties in feature extraction and feature classification in underwater object sonar image recognition are discussed. Finally, a detailed discussion of underwater object tracking methods and algorithms based on acoustic signal processing and sonar image information is given. By the indepth discussion of every underwater target dealing process, analysis and comparison, some key scientific issues and possible solutions of sonar-image-based underwater object detection, recognition and tracking are summarized, which is followed by a forecast of future development directions in this field.