In dynamic environments, mobile robots sometimes are required to perform accurate and stable tracking of high-speed trajectories, but rare researches are reported, which specifically deal with high-speed tracking problem. When tracking high-speed trajectories, mobile robots usually approach the top speed and acceleration limits, which indicates that the kinematic constraints can not be ignored. In order to cope with such situations, a novel trajectory tracking controller using model predictive control has been studied. Where it differs from other control strategies lies in that this control law is generated based on predicating the evolution state of mobile robots and it is able to handle hard constraints directly in the optimization process, so that mobile robots can track trajectories both quickly and safely. In order to cut down the computational cost for on-line applications, Laguerre polynomials are used in the design of model predictive control to reduce the number of parameters for optimization. The proposed algorithm is applied on a real soccer robot, and the experimental results show that the robot can track high-speed trajectories effectively with small tracking errors and tiny computation time.