According to the wind speed and output power historical data of wind turbine, the relationship between the average output power of the wind turbine and the wind speed was subjected to least squares parameter identification, and the relationship between the output power deviation of the wind turbine and the wind speed was obtained based on the above research. The Python probabilistic analysis was used to determine the probability distribution types of each subset of the wind turbine output power deviation, and the characteristic parameters were estimated to obtain the probability density function, furthermore, the probability of confidence interval prediction was performed on the wind turbine output power deviation. Based on the relationship between the average output of the wind turbine and the wind speed and the confidence interval estimation model of the wind turbine output power deviation subset, the confidence interval of the wind turbine output power was predicted according to the wind speed forecast value. The method proposed was tested and verified with the historical record of the actual wind turbine, the confidence interval of the wind turbine output power was predicted according to the wind speed forecast value. The results showed that the wind turbine output power deviation was divided into multiple subsets based on the wind speed, which could improve the accuracy of the confidence interval prediction of the wind turbine output power.