This paper adressed the problem of state and road adhesion estimation by using the Takagi-Sugeno (TS) approach. A Lateral vehicle model is used with nonlinear contact forces. From the considered model an exact TS model with unmeasured premise variables is deduced by the well known sector nonlinearity transformation. In this model, the road adhesion coefficient is expressed as an unknown input which is also involved in the weighting functions of the TS model. An observer is then designed for state and road adhesion estimation which is an important parameter in vehicle safety systems. The convergence of the observer is studied with Lyapunov theory and conditions that ensure Input-To-State Stability (ISS) are provided and expressed in Linear Matrix Inequalities (LMIs) Some simulation results are provided with a test on real data to illustrate the performances of proposed approach.