Asimulation before test method for fault diagnosis and classification towards sensor fault in linear time invariant state feed back system is presented in this paper. The novelty of the approach lies in associating with each state feedback gain factor a scalar, which is defined as the sensor healthiness factor. This scalar is made to vary from 1 (no fault condition) to 0 (full fault condition) in predetermined steps. The intermediate values of portray the deterioration modes of the sensor. The Integral Square Error (ISE) criterion is employed for extracting the signature of the fault and the classification is done using Artificial Neural Network (ANN) classifier. The proposed diagnosis approach is applied to a dc motor system to validate the effectiveness of the technique.program inspections, static & dynamic analysis and V&V techniques

Key words : Neural Network, Integral square error, fault diagnosis