Since many wind turbines are installed at remote locations, the introduction of fault diagnosis and fault-tolerant control is considered a suitable way of improving reliability of wind turbines and lowering costs of repairs. In this project, a 4.8 MW fictitious but realistic wind turbine is considered, for which a dynamical model is derived. A fault analysis is conducted to identify the frequency of occurrence and the severity of the end-effects of possible component faults. Methods for diagnosis and accommodation of the most significant faults are then applied. The diagnosis algorithms are based on a common structure relying on a reconfigurable extended Kalman filter, which allows diagnosis of multiple simultaneous faults. Generally, the abrupt faults are diagnosed using hypothesis testing based methods, while the incipient faults are diagnosed using parameter estimation based methods. The fault diagnoses algorithms can be used for both conditioning monitoring and active fault-tolerant control purposes. Fault-tolerant capabilities are obtained by correcting the faulty signals or by incorporating fault-tolerance in the control system. Both active and passive fault-tolerant control systems are designed based on LPV methods, due to the parameter-varying nature of the wind turbine, and are compared in terms of design complexity and performance. Verification of the control systems confirm that they are capable of controlling the wind turbine exposed to multiple simultaneous faults; consequently, the reliability of wind turbines can be improved.