This paper studies the application of estimation error signals frequency domain analysis in the fault detection field. The estimation error signals of actuators are derived from the improved set-valued observer and the benchmark mathematical model of a wind energy conversion system (WECS) with delayed inputs and unknown nonlinear parts. The improved SVO is designed based on the extended Lipschitz condition, the set-induced theory, and the uniformly ultimately boundedness. For the faults of the pitch angle actuators, the estimation error signals are processed by the fast Fourier transform (FFT), and the corresponding fault detection strategies are designed. The simulation results show that the proposed fault detection strategy can detect the actuator fault in the early stage of the fault with high accuracy.