The classical model-reference adaptive control approach works satisfactorily under certain assumptions on the unknown process only. When these assumptions are not satisfied, the resulting control system is unstable, and an alternative to obviate this pitfall is to use fuzzy adaptive control. Here the model of the plant is defined by the fuzzy Takagi-Sugeno model, and the algorithm of the globally stable fuzzy adaptive control is based on the Lyapunov criterion. The main idea of the approach is the fuzzification of the adaptive parameters corresponding to inputs, outputs and state variables. The paper deals mainly with systems involving nonlinearity. Two simulation examples are displayed.