Abstract: A robust tracking controller, combing an extended state observer (ESO) and adaptive backstepping control technique, is designed to solve the unavoidable uncertainties of unmanned helicopter in real life flight. Firstly, the mathematical model of the unmanned helicopter is established. Different types of uncertainties such as the unmodeled dynamics, model simplification error, external atmospheric disturbances and inertia parameter perturbation, caused by the gradual fuel consumption and so on, are taken into account during modeling. In addition, the first order flapping dynamics of the main rotor is coupled into the 6DOF rigid body model of the unmanned helicopter, establishing a simple equivalent model but reflecting the unique characteristic of the flapping dynamics of the main rotor. Then, the position controller, attitude controller and torque controller are designed based on the ESO and the backstepping method. The adaptive control strategy is applied to estimate slowly varying perturbations of the parameters such as unmanned helicopter mass and inertia matrix, while the ESO is used to observe the high-frequency disturbances such as the unmodeled dynamics and external gust. And the feedforward compensation is carried out subsequently in the control law to realize the comprehensive suppression of the uncertainties of the different disturbing frequencies in the unmanned helicopter system. Finally, the feasibility and validity of the method are verified by numerical simulation. The simulation results show that the method can eliminate the unmodeled dynamics and uncertainties of the unmanned helicopter, and has good robust performance. The simulation results imply that the proposed method has higher uncertainsuppression efficiency and control accuracy than the traditional adaptive backstepping method, and can realize the robust tracking control of the unmanned helicopter trajectory, under the situation when many different kinds of disturbances act together.