In recent years, the frequent occurrence of extreme weather has greatly affected the safe and stable operation of the power grid. When the ice and wind are coupled, the transmission line are prone to cause large-scale power failure. Based on the analysis of typical ice-wind fault mechanism of the transmission line, this paper simulates the relationship between the wind speed and the wire tension under the ice-covered condition, and introduces Naive Bayesian method, which shows good performance in small samples, to analyze the ice-wind fault. In order to improve the performance of Naive Bayesian method, this paper proposes a Naive Bayesian classification model based on correlation analysis, in which the Spearman correlation analysis method is used to extracted typical parameters as the input of the model. To verify the good performance of the proposed model, fault cases of a power grid company in the last five years are used and the results show that the proposed model outperforms the original Naive Bayesian method, yielding average classification accuracy of 99.24% in the identification of ice-wind fault, which provides a strong support for the follow-up research of ice-wind fault identification.