This paper studies the model based fault tolerant tracking control problem for continuous-time nonlinear systems that are represented by the Takagi-Sugeno fuzzy models with multiplicative noises. The multiplicative noise terms are introduced in the consequent part of the Takagi-Sugeno fuzzy system to represent the stochastic behaviors. The faults considered in this paper are characterized as time varying functions that are modeled by exponential functions or first order polynomials. The purpose of this paper is to ensure system states can track the state references when faults occur. Based on Lyapunov theory and descriptor redundancy property, the sufficient conditions are derived in terms of linear matrix inequalities. Finally, some numerical examples are provided to illustrate the applicability and effectiveness of the proposed control methodology.