A design method for the extension controller is developed in this paper. The proposed adaptive extension control resulting from the direct adaptive approach is employed to directly adapt the gain parameter of the extension controller. Then the constructed controller can be best approximated to a given optimal control. Unlike the fuzzy controller, only one linguistic — like level is needed in the extension controller. The merits of the proposed controller are that (a) the number of adaptation parameter is small; (b) the design algorithm is easily to be implemented. In addition, a maximum control is established to guarantee the system robust stability. The derivation shows that the proposed extension controller is stable in the sense of the Lyapunov. Finally, a nonlinear system simulation example is applied to verifying the effectiveness and the ability of the proposed adaptive extension controller.