Abstract: In order to meet the requirements of large-scale underwater audio synchronous monitoring, a comprehensive scheme of distributed system clock synchronization and underwater acoustic monitoring was proposed. IEEE1588v2 PTP protocol and GNSS timing synchronization technology were used for distributed clock synchronization. In terms of software development, the underwater acoustic monitoring software is developed, and the underwater acoustic recognition is carried out by convolutional neural network model. After evaluation and testing, according to the scheme, the whole system clock synchronization accuracy based on satellite timing reaches 20µs, the power consumption is less than 10W, and the Marine audio recognition is completed based on the TensorFlow Lite convolutional neural network model with the size of 4M and the speed of 2-3 s. The monitoring platform can combine multiple nodes for underwater audio monitoring, simultaneously collect underwater acoustic data synchronously and process necessary algorithms. Moreover, the algorithm can choose traditional processing algorithm or deploy convolutional neural network for more complex functions, so it has broad development prospects.