Abstract: Under more general conditions, the stability and the robustness of goldensection feedback control law are analyzed in this paper, and in the framework of hybrid identification,  the inputstate stable(ISS) property of the gradient project method used in  SISO and MIMO characteristic models is studied, and  the L2bound of estimated errors of characteristic parameters is obtained. The final results further generalize the former results and play very important roles in the stability analysis of entire hybrid system.

Key words: allcoefficient adaptive controller, characteristic model, golden section, gradient project method