The purpose of this paper is to propose a systematic smooth control method for improving the stability of the two-wheeled self-balance robot under impact disturbances. For enhancing the robtot stability, a blend controller based on states feedback control embedded with the PID speed synchronization is estabilished, as well as a hybrid filter composes of RC network and Kalman algorithm. With the hybrid filter, disturbance signals are maximally attenuated or directly eliminated, and the system sensitivity to the impact disturbances significantly declines ; under the blend motion controller, the robot can not only keep balance under impacts but also achieve synchronization of the two driving wheels. The dynamic model, the blend controller, hybrid filter, and experimental results including application to transport are described, both of the simulation and experimental results are provided to verify the analysis.