The presence of cracks on the surface of MEMS microelectromechanical system (MEMS) devices affects their functional parameter, such as resonance frequency. Overtime, these cracks may cause the devices’ failure. Therefore, it is important to identify the symptoms of the cracks presence so that the affected devices can be removed/bin-out before delivering to the customers thus avoiding failure during applications of these devices. In this paper, we present a predictive and quantitative fault modeling approach using the linear time invariant (LTI) technique for a piezoelectric cantilever-based sensor. The analytical fault model is developed by correlating single edge, double edge, and surface cracks with the resonance frequency of the cantilever. The Simulink is utilized to design the LTI-based faulty device by integrating the dynamic equations associated to actuation and sensing mechanisms of the device under study. The model is validated by comparing it with the finite element method model containing the same design parameters, using the COMSOL Multiphysics. We observed that the presence of cracks at any location on the surface of the device stimulates the amplitude that causes decrease in its resonance frequency. The increase in the amplitude of vibration due to the presence of cracks is electrically sensed by the piezoresistive mechanism. Finally, the fault detection methodology is presented by using the Stateflow technique. We found that the proposed techniques can be significant to study the faulty behavior of the MEMS devices.