Abstract: A new fuzzy adaptive hierarchically triggered control method for muscle strength training is presented to solve the deficiency that the muscle strength training methods in existing robot-aided resistance training systems are all constructed within the framework of predefined reference trajectory by designing low-level resistive force controller according to the participant's motor behavior. The new method is developed without the need of predefined training trajectory, and firstly a high-level progressive resistance supervisory controller of resistive force is designed based on the impaired limb's motor performance to determine the basic resistive force for each training session. Secondly, a low-level adaptive resistive force triggered controller is constructed according to the impaired limb's bio-impedance changes to further adjust the resistive force in each training session. Finally, the effectiveness and potentialities of the proposed control strategy are verified with clinical experimental results.