A mobile data collection scheme based on compressive sensing is proposed in order to reduce the energy consumption and ensure real-time performance of data collection in UnderwaterWireless Sensor Network(UWSN). Firstly, based on the Distributed Energy-Balanced Unequal Clustering (DEBUC) protocol and compressive sensing theory, cluster nodes decide whether to participate in compressive sampling according to the designed sparse measurement matrix, and transfer the obtained measurement results to the cluster head. Then, the data at cluster head is mobile collected by Autonomous Underwater Vehicle(AUV) to the data center. This problem is modeled as a Traveling SalesmanProblem(TSP) based on the maximization of information quality, and an approximate algorithm is proposed to solve this problem. The simulation results show that, compared with the existing underwater data collection algorithms, the proposed scheme effectively reduces the data collection delay and prolongs the lifetime of network, with assuring the reliability of data collection.