This paper presents how Pariially Observable Markov Decision Processes (POMDPs) can be used for controlling fleets of VAS under uncertainty l. POMDPs provide a soundmathematical framework to deal with planning actions when tasks outcomes and perception are uncertain, although their computational complexity have precluded their use for multi-robot applications. However, in this work a scalable approach based Oil POMDPs is presented. Instead of solving a complex model for the whole team, distributed models in which the VAS have no knowledge about others' actions are considered. Then, a decentralized data fusion algorithm is used to share information from all the sensors and obtain an implicit coordination between the VAS. Moreover, the proposed approach is applied to a tracking application with a fleet of VAS, and some simulations are presented to show its advantages and how the mentioned coordination arises between the members of the team.