Abstract: Wind power gearboxes work under variable speed conditions. The vibration signal of the gearbox has the characteristics of being multi-component modulation and the characteristic frequency of the fault being subjected to frequency modulation. Aiming at the problem of difficult identification of gearbox fault characteristics caused by this feature, a method for fault diagnosis of wind turbine planetary gearbox based on order analysis of speed estimation is proposed. Select the sensor measurement points by analyzing the structure and fault characteristics of the wind turbine gearbox, and complete the selection of the sensor according to the gearbox vibration frequency range, build a monitoring system on the basis of the above work, and collect the vibration of the gearbox under working conditions Information, and perform rotation speed estimation and order analysis on the vibration data collected on the actual site to extract the order of fault characteristics of the gearbox and complete the fault diagnosis. This method has been verified by the wind farm, and the fault diagnosis accuracy rate reaches 84%, which is 8% higher than the traditional FFT spectrum analysis method, which can meet the actual engineering needs of the site.