A GA-based fuzzy controller design method is proposed for a two-wheeled mobile robot to independently control two velocities of its left-wheeled motor and right-wheeled motor so that the controlled robot can move fast and efficiently to the desired position. First, the kinematics model of the considered two-wheeled robot is described and a 4-input-and-2-output motion fuzzy control structure with two 2-input-and-1-output fuzzy systems is proposed. Then a GA-based method is proposed, where a ratio coefficient coding method and a variable fitness function in the genetic algorithm are proposed to choose the input and output membership functions of these two fuzzy systems so that the selected controller has a good performance in the robot's motion. From the control results of the proposed method in the 3D Robot Soccer Simulator of FIRA, we can see that the proposed motion control method for two-wheeled mobile robots is feasible and effective.