The control problem for a class of complex systems with the disturbance of white noise and additivity is
considered. The Takagi-Sugeno fuzy control model is established, the local state information is estimated by Kalman filter,
and the control gain is obtained by dynamic programming. Then the controller is designed, which has the active learning
feature and the desired performance. Using an inverted pendulum as a simulation example, the results show the effectiveness
of the designed controller.