A modified Strong Tracking Filter (STF) is used to develop a new approach to sensor fault tolerant control. Generic Model Control (GMC) is used to control the nonlinear process while the process runs normally because of its robust control performance. If a fault occurs in the sensor, a sensor bias vector is then introduced to the output equation of the process model. The sensor bias vector is estimated on line during every control period using the STF. The estimated sensor bias vector is used to develop a fault detection mechanism to supervise the sensors. When a sensor fault occurs, the conventional GMC is switched to a fault tolerant control scheme, which is, in essence, a state estimation and output prediction based GMC. The laboratory experimental results on a three tank system demonstrate the effectiveness of the proposed Sensor Fault Tolerant Generic Model Control (SFTGMC) approach. A modified Strong Tracking Filter (STF) is used to develop a new approach to sensor fault tolerant control. Generic Model Control (GMC) is used to control the nonlinear process while the process runs normally because of its robust control performance. in the sensor, a sensor bias vector is then introduced to the output equation of the process model. The sensor bias vector is estimated on line during every control period using the STF. The estimated sensor bias vector is used to develop a fault detection mechanism to supervise the sensors. when a sensor fault occurs, the conventional GMC is switched to a fault tolerant control scheme, which is, in essence, a state estimation and output prediction based GMC. the proposed Sensor Fault Tolerant Generic Model Control (SFTGMC) approach.