The main bearing of wind turbines is an important part of the transmission system, and its anomaly identification accuracy is greatly affected by wind speed fluctuations. To solve this problem, an abnormal identification method for the main bearing of wind turbines based on BPNN-NCT is proposed. First, the correlation coefficient method is used to determine the parameters related to the main bearing state as the input of the model, and a state parameter prediction model with the main bearing temperature as the state parameter is established based on a back propagation neural network (BPNN). Then a non-central t (NCT) distribution is used to describe the distribution characteristics of the state parameter prediction residuals under different wind speed fluctuation intervals. Then a quantitative index of the abnormal state of the main bearing considering the influence of wind speed fluctuations is proposed. Finally, a 2 MW direct-drive wind turbine of a wind farm is taken as an example to verify the accuracy and effectiveness of the proposed method. This work is supported by the National Natural Science Foundation of China (No. 61379063).