A mobile manipulator can perform various tasks efficiently by utilizing mobility and manipulation functions. The coupling of these two functions creates a particular kinematic redundancy introduced by mobility, which is different from that introduced by extra joints. This redundancy is quite desirable for dexterous motion in cluttered environments, but it also significantly complicates the motion planning and control problem. In this paper we propose a new motion planning method for mobile manipulators to execute a multiple task which consists of a sequence of tasks. The task considered in this paper is that the end-effector tracks a prespecified trajectory in a fixed world frame. In a multiple task, the final configuration of each task becomes the initial configuration of the next subsequent task. Such a configuration is known as a commutation configuration, which significantly affects the performance of the multiple task. We formulate the motion planning problem as a global optimization problem and simultaneously obtain the motion trajectory set and commutation configurations. In the formulation, we take account of the case that the platform has a non-holonomic constraint as well as the one that the platform has a holonomic constraint. Simulation results are demonstrated to verify the effectiveness of the proposed method.