Abstract: Accurately depict the short-term wind power prediction error and the characteristics of regional wind power output fluctuation is solved,the basis of large-scale parallel operation problem of uncertainty in energy for accurate characterization of wind power output fluctuation and the prediction error and the error of meteorological correlation,Gaussian mixture distribution probability model is set up and use the cloud model and the error of the observation curve structures,Then based on the coupling of the normal cloud and mixture Gaussian distribution probability distribution model,finally using a variety of probability density distribution model of single wind power prediction error,the cluster wind power prediction errors,the region’s weather errors as well as different power range of northern Hebei Province weather prediction error of the prediction error and the corresponding statistical analysis of correlation. The simulation results show that the proposed model has the best fitting effect,which verifies the effectiveness of the probabilistic model based on the coupling of normal cloud and mixed Gaussian distribution.