Aiming at the problem that the reliability of key components such as generator and gearbox of wind turbine is difficult to be accurately evaluated due to the bad working environment of wind turbine, a reliability evaluation method based on improved weibull distribution model is proposed by taking the gearbox of a certain type of wind turbine as an example. This method optimizes the traditional mean rank method, and overcomes the problem that the traditional mean rank method only considers the location of each non-outage fault, but does not consider the influence of the length of outage fault interval on the results, resulting in parameter estimation error. The weibull distribution model and the improved weibull distribution model are used to evaluate the key components of wind turbine based on the outage and non-outage data of the key components of wind turbine. The reliability function curve and failure rate function curve are obtained and analyzed. Through case analysis and comparison of parameter estimation results of the two models, it is verified that the proposed new method can calculate parameters with higher accuracy, making the evaluation results more in line with the actual situation of wind turbine operation.