This paper develops a fuzzy model reference adaptive control of continuous-time nonlinear systems described by Takagi-Seguno (TS) fuzzy model. The tracking performance is assured using a direct TS fuzzy controller. Both full state information and observer-based control schemes are investigated. It is proved, using Lyapunov stability tools, that this adaptive scheme is asymptotically stable and the tracking and observation errors converge to zero. Simulation results illustrate the performance of the proposed approach.