Abstract: Based on chaos variables, an optimal design scheme for of fuzzy controller is proposed in the paper. A chaos algorithm is used in off-line optimization, and chaos factors are incorporated into optimized searching of fuzzy controller parameters. The optimal variables are transformed into chaotic variables by carrier wave. Then, we take advantage of ergodicity and randomness of chaos dynamics in the searching process, and get the approximate global optimal values of the parameters. A conjugate gradient descent algorithm is used in on-line optimization, and the initial values of gradient descent searching are used as approximate global optimal values of chaos searching, and controller parameters are quickly optimized by means of combination of chaos global searching and gradient descent local searching. Finally, convergency of the algorithm is proved.