This paper considers the path planning problem for deployment and collection of a marsupial vehicle system which consists of an unmanned ground vehicle(UGV) and two flying robots. The UGV, as a carrier, is able to deploy and retrieve the flying robots while each of the latter usually takes off and lands on the carrier. This work is motivated by cooperative search and reconnaissance missions in field of heterogeneous robot system. A sequence of tasks, whose positions are known as prior knowledge, are expected to be accessed by the flying robots in a given order. Owing to their limited duration in the air, the aerial robots, usually unmanned aerial vehicles (UAVs), should return to the UGV timely and land for its energy-saving and recharge. This paper mainly contributes on the mathematical model for path planning of the special marsupial system. Many real constraints including the maximum velocity of the UAV and the UGV, the maximum hovering time and so on are taken into account in the model. The objective function is constructed by minimizing the time for completing the whole mission. In order to clarify the novel method, a single task point model first is proposed and then extended to multi tasks one. Finally, particle swarm optimization (PSO) method is used to solve the path planning problem. Several simulations verify the feasibility and effectiveness of the proposed method.