As the technologies for single unmanned systems have matured to a level of practical applications, the interest of cooperative operation with multiple unmanned systems has been rapidly increasing. This study focuses on developing mission planning algorithms for search and surveillance in a marine environment using multiple heterogeneous unmanned systems composed of unmanned surface vehicles (USVs) and unmanned aerial vehicles (UAVs). The roles and the mission planning architecture are determined considering the specifications and sensing capabilities of the vehicles. In particular, a constraint-based task allocation algorithm is proposed as an essential part of the mission planner and its validity is demonstrated using numerical simulations with various search scenarios.