This paper presents a parallel control method based on the expanded state observer (ESO) for aero-engine blade robot polishing. Aiming to reduce the fluctuation of polishing force caused by environmental noise and modeling errors. First, calibrate six-dimensional force sensor according to the maximum acceleration of the end effector during the polishing process. Then, build the gravity compensation and zero drift compensation model. Besides, use this model to compensate measurement error of the six-dimensional force sensor. Finally, calculate the error between the expected polishing force and the actual feedback value and its derivative value. Use calculation results to design the control boundary layer. The polishing force controller is divided into two parallel control loops to design. When the switching value is in the control boundary layer. A nonlinear active disturbance rejection control (ADRC) loop is used. When the switching value is outside the control boundary layer. An ESO-based sliding mode control (SMC) loop is used. Simulation and experimental results show that the proposed parallel control method based on ESO has a fast response and high robustness compared with FuzzyPID, PID, and ADRC. It can effectively suppress the force fluctuation in the polishing process and significantly improve the surface processing quality of the aero-engine blade.