As wind energy plays an increasingly important role in the US and world electricity supply, maintenance of wind turbines emerges as a critical issue. Because of the remote nature of wind turbines, autonomous and robust health monitoring techniques are necessary. Detecting faults in complex systems such as wind turbine gearboxes remains challenging, even with the recently significant advancement of sensing and signal processing technologies. In this paper, we collect time domain signals from a gearbox test bed on which either a healthy or a faulty gear is installed. Then a harmonic wavelet based method is used to extract time-frequency features. We also develop a speed profile masking technique to account for tachometer readings and gear meshing relationship. Features from multiple sources are then fused together through a statistical weighting approach based on principal component analysis. Using the fused timefrequency features, we demonstrate that different gear faults can be effectively identified through a simple decision making algorithm.© (2011) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.