In this paper, a sensorless pitch angle control method for a wind generation system is suggested. One-step-ahead prediction control law is adopted to control the pitch angle of a wind turbine in order for electric output power to track target values. And it is shown that this control scheme using the inverse dynamics of the controlled system enables us to predict current wind speed without an anemometer, to a considerable precision. The inverse input-output of the controlled system is realized by use of an artificial neural network. The proposed control and wind speed prediction method is applied to a Double-Feed Induction Generation system connected to a simple power system through computer simulation to show its effectiveness. The simulation results demonstrate that the suggested method shows better control performances with less control efforts than a conventional Proportional-Integral controller.