Abstarct: In this research, first a comprehensive dynamic model for the Wind Energy Conversion Systems (WECSs) is presented by which a variety of faults such as sensors, actuators and power grid faults can be studied and effects of these faults on voltage, current and other signals of the WECS can be investigated. Many modelling techniques which have been already introduced for the modelling of the faults or observing the influences of the faults cannot be considered wide-ranging to include any fault at any point of the WECS. Then, a Fault Detection and Isolation system (FDIS) is presented to diagnose the occurred faults at the appropriate time in order to ensure a safe system operation, avoid heavy economic losses, prevent damage to adjacent relevant systems and facilitate timely repair of failed components. The proposed FDIS, which exploits dynamic recurrent neural networks, detects and isolates the occurred faults in the generator's angular velocity sensor, pitch angle sensors and pitch actuators. The novelties of the thesis can be explained as follows: the comprehensive dynamic model has been applied to analyze faults of the WECS and to design the corresponding FDIS. Also utilizing the mean of the residual signal has been proposed to isolate the faults of the pitch system. Then the adaptive fuzzy threshold has been presented to achieve to the robust design. Comparison of the simulation results for the constant, adaptive, and adaptive fuzzy thresholds shows that the numbers of missed and false alarms have reduced by using the adaptive fuzzy threshold. Finally by employing the designed FDIS, a Fault Tolerant Control System (FTCS) is suggested which contains both the signal modification and the bank of controllers blocks. Also when a fault has occurred in the angular velocity sensor of generator, if the suggested FTCS is utilized, the performance of the WECS will be raised and more active power can be attained. The suggested FDIS and algorithm can be used to detect occurred faults in other parts of the WECS.