The invention discloses a sliding-mode self-adaptive control system of a two-wheeled self-balancing robot. The sliding-mode self-adaptive control system comprises a sensor measurement module, a main control chip and a motor system, wherein the sensor measurement module is used for collecting motion parameters of the self-balancing robot, and the motor system is used for driving the two-wheeled self-balancing robot to move; the main control chip is connected with the sensor measurement module and the motor system and used for controlling motions of the motor system according to the motion parameters collected by the sensor measurement module; and a sliding-mode self-adaptive controller is arranged in the main control chip, and the output equation of the sliding-mode self-adaptive controller is U=-(K+phi)*X. By adopting the technical scheme, while the robot is self-adaptive to the external environment, the influence of various interference factors in the external environment on the two-wheeled self-balancing robot can be lowered to the maximum extent, and the robustness is not lost; and meanwhile, the two-wheeled self-balancing robot has the optimal performance through running-in of long-term accumulated factors by means of machine learning, and accordingly the safety and stability are guaranteed.