This paper designs a sensor-based behavior fusion mechanism for car-like mobile robot to implement the autonomous driving mission. First, a design of collision prevention system is introduced for the car-like mobile robot. Secondly, the wall-following controller composed of a back propagation neural network and a fuzzy logic control is proposed to make the CLMR move smoothly in the unknown and changing environment. Moreover, the controller design for auto-parking issue which is not only considered the normal parking condition, but taken the complex parking condition into account, is also revealed in this manuscript. Computer simulation results illustrate the effectiveness of the proposed sensor-based behavior fusion mechanism. Finally, the real-time experiments of the autonomous driving control on the test ground demonstrate the feasibility of CLMR maneuvers.