This paper is concerned with the problem of fault tolerant control for non-linear systems with sensor faults based on the T-S fuzzy model approach. First, a fault estimation (FE) observer with (Formula presented.) technique is designed to estimate the fault. Second, by using the fault estimation information, which are provided by the fault estimation observer, a novel fault tolerant control scheme with a fault compensation mechanism in the premise part of the fuzzy controller is presented to compensate the effect of the fault and stabilize the system. Different from the existing methods, where the fault compensation mechanism is only established in the consequent part of the controller, in the presented method, the fault compensation mechanism is established in the consequent part as well as the premise part of the fuzzy controller. Then, by using the fault compensation mechanism, the bound of the deviation between the normalized membership functions of the fuzzy model and those of the controller can be reduced such that the fault tolerant control design can be relaxed. In the end, an example is given to illustrate the validity of the developed fault tolerant control scheme.