Abstract: To diagnose the faults of a wind turbine blade pitch actuator more quickly and accurately, an adaptive linear parameter varying (LPV) fault observer based on the established LPV model of wind turbines is designed. Limiting conditions of the adaptive LPV fault observer are obtained by considering the influence of the wind disturbance and faults on the stability of the system and extending fault observer constraints to the LPV systems. The LPV system could be approximated by adapting the convex decomposition technique. The problem of solving infinitely many inequalities has now been translated into the problem of solving finite inequalities. The observer gain and the fault estimation update gain that vary with the scheduling variables are solved by using linear matrix inequality tools. Simulation results demonstrate that compared with the adaptive fault observer algorithm, this algorithm has better rapidity and sensitivity in the treatment of wind turbine pitch actuator faults under sinusoidal fault, high content of air hydraulic, and pump blockage fault.