This thesis enhances the indirect adaptive fuzzy control approach in following directions: The proposed approach is less conservative concerning the stability of the plant zero dynamics and allows an optimal controller design, the attenuation of measurement noise and the inclusion of linear observers. An automatic control for a class of nonlinear uncertain MIMO systems is developed, which requires only less information about the controlled process. However, if there is any linguistic process model description from experts available this knowledge in form of fuzzy if-then rules can be included during the controller design. The robustness and performance of the proposed tracking control will be shown based on fuzzy model error compensation and the attenuation of external disturbances as well as measurement noise in their impact on the control error. A Lyapunov stability proof is stated which guarantees all included signals to be bounded and the fuzzy parameters convergence to their optimal values. Appropriate projection algorithms were chosen to restrict the parameters of the adaptive fuzzy systems within constraint sets. A SPR Lyapunov design approach allows the inclusion of linear state observers in the control concept and a dynamic fuzzy rule activation method remedies the phenomenon which is called the curse of dimensionality. Simulations of an inverse pendulum system and a magnetic levitation system were carried out to show the practicability and high performance of the proposed approach. A short discussion of the qualities and limitations of the proposed method concludes the thesis.