The vibration monitoring method is the main approach of condition monitoring for wind turbine transmission system. The vibration monitoring strategy and the process of each component′s vibration feature extraction of wind turbine transmission system are analyzed firstly, and the trend indicators of vibration feature such as side band power factor (SBPF) and side band energy ratio (SER) are introduced especially. Then, it is pointed out that the key to retrieve the huge economic losses caused by the fault of wind turbine transmission system is incipient fault prediction for wind turbine transmission system, and the generalization manifold learning-based incipient fault prediction method for wind turbine transmission system is introduced especially. Finally, the function and characters of the existing vibration monitoring system for wind turbine transmission system are analyzed including system architecture, data acquisition configuration and monitoring analysis methods. Furthermore, it is pointed out that big data-based predictive analytics and intelligent maintenance based on multi-source information fusion technique would be the important development trend of health management for wind turbine transmission system.