Abstract: This paper presents a new approach to automatic tuning of fuzzy logic controller for manipulators, which modifies the shape of membership functions automatically according to previous responses to achieve a desired control performance. On the basis of defining a set of cubic splines as membership functions regarding change-in-error ej(t), the optimal membership functions are found by the Nelder and Mead's simplex algorithm. The proposed control scheme consists of a traditional fuzzy logic controller and a conventional derivative controller. In order to demonstrate the effectiveness and the robustness of the control scheme, we report a number of simulation results of a two-link manipulator in the presence of load and friction changes.