Abstract: At present, wind turbine application fault warning is mostly based on the big data cloud platform mode, and wind farm station managers cannot timely and effectively grasp whether wind turbine operation risks exist and can only passively receive fault warning results from the cloud and passively carry out on-site operation and maintenance according to operation and maintenance suggestions. Therefore, a wind turbine fault warning platform is designed for deployment and application of wind farm stations. Firstly, the overall design of the platform is described from the platform framework and function module design. Secondly, combined with the background of parity online in the wind power industry, a general design method of fault early warning based on anomaly deviation monitoring is proposed based on the analysis of existing fault early warning methods, and the realization of the fault early warning function of the platform is emphasized. Finally, in order to verify the practical application effect of the platform, the pilot wind field is used as a test sample to deploy the platform. The actual operation of the wind farm shows that the platform system is stable and reliable and can meet the practical needs.