In this study, a hierarchical fuzzy controller associated with PD classical controller with feedback error learning method for class of canonical SISO nonlinear system in presence of bounded disturbance is presented. The stability of whole system is guaranteed through a Lyapunov function. The adaptation laws of all parameters of consequent part of hierarchical fuzzy system are derived using it. Tunable parameters of hierarchical fuzzy system are appeared in nonlinear form at the output. Using the mentioned theorem, they are appeared in linear form. There is an upper bound for the residual terms. They are considered in Lyapunov function and the adaptation law is derived for them. It is shown that, the proposed theorem is applicable to the hierarchical fuzzy system with any structure and any number of layers and even to the ordinary fuzzy systems. Also, using hierarchical fuzzy controllers leads to reduction of number of rules and parameters in a fuzzy system. Finally, the proposed method is applied on two systems (flexible joint robot and quarter active suspension system). The results are compared with classical sliding mode method. They reveal the efficiency of the suggested algorithm.