Abstract: When the oil temperature of the gearbox is too high, the wind turbine will lead to the limited power operation of the unit, which will affect the generating efficiency of the unit. Threshold judgment is often used to judge the wind turbine’s high temperature capacity reduction state, which results in slow response and aggravates the deterioration trend of wind turbine gearbox. In this paper, Bayesian network was used to evaluate the wind turbine’s high temperature capacity reduction state. In order to extract and make accurate and reasonable use of the coupling characteristics between the various state parameters of the supervisory control and data acquisition system（SCADA）, the state parameters of the unit were introduced by vine-Copula model. Through correlation analysis, Bayesian probabilistic graph network which is more in line with the actual operation state of the unit was established to realize the evaluation of the unit’s high temperature capacity reduction state. The cross-entropy algorithm was used to evaluate the output of the model. It is found that compared with the naive Bayesian model, the vine-Copula Bayesian network is more accurate and reliable, and the model is more in line with the actual operating conditions of the unit. It can provide a reference for the operation and maintenance personnel in the field to formulate accurate and reasonable operation and maintenance schemes.