On the basis of analyzing characteristics of the central air-conditioning system, a fuzzy controller with self-optimizing parameters is proposed and applied into the temperature control of the system. To deal with the non-linearity, time delay and time variation of the controlled system, a method for optimizing the weighting factor a and the output scaling gain K_u is adopted in the applied rule-adjustable fuzzy controller. The experimental result indicates that the optimization method enhances the robustness of the fuzzy controller. As the temperature of the cooling water and the temperature of the circumstances change, the controller maintains satisfactory performance.