Abstract: Based on the variation analysis of electric network loads on various holidays, the models for forecasting the daily load of the electric power network and its maximum and minimum loads per day in the Central China are given.Using meteorological variables as predictor, the dynamical colligating linear regression combined with auto-regression and the dynamical nonlinear artificial neural network methods are used.The prediction with 365 independent samplesshows that the methods have a high precision in electric load forecasting in Central China and can be used for the operational practice.