This paper describes an investigation into the coordination of multiple autonomous underwater vehicle system (MAUVS). In order for MAUVS to work efficiently in such scenarios as oceanographic data gathering, a market-based approach to coordination of multiple underwater vehicles is proposed. In this approach, robots on the team act as self-interested agents and a virtual economy environment in which tasks and resources are exchanged based on one's own benefit was established. While robots trade tasks and resources with one another to maximize their benefits, the efficiency of the whole system is reached. Market based coordination of MAUVS is further integrated into MOOS (Mission Oriented Operating Suite-an open source project proposed and maintained by Newman from Oxford and Benjamin from MIT) to enable the MAUVS to work efficiently and intelligently. Simulation results demonstrate the efficiency of the approach proposed in this paper.