Abstract: In order to meet needs for electric power of meteorological service, an analysis is made on the correlation between meteorological elements and electrical loads of electric network in Hebei Province. The meteorological data and electrical load data from May to September during 2001-2010 are used, and they are divided into sunny hot weather and muggy weather. Compared to the sunny hot weather, it shows that the daily peak load and daily valley load are increased significantly in muggy weather lasting three days or more. When daily maximum temperature reaches 32℃, daily peak load of power increase rapidly in Hebei Province, and 32, 35℃ and 38℃ of daily maximum temperature are three sensitive points for daily peak load of power to air temperature variation. During periods with daily maximum temperature more than or equal to 35℃, the daily peak load of power varies greatly according to the air temperature. When daily maximum temperature exceeds 38℃, considering 1℃ rising of daily maximum temperature, the daily peak load of power would increase 9.4%, and the air-conditioning cooling load of power would reach 50% of the daily peak load. When daily minimum temperature reaches 25℃, daily valley load of power increases rapidly, and 25℃ of daily minimum temperature is the sensitive point of daily valley load of power to air temperature variation. Introducing accumulated temperature effect as forecast factor, a meteorological electricity prediction model is established by using the multiple regression method, which can predict the peak and valley of meteorological electricity loads. The model is validated using historic data from 2011 to 2013, the average relative error of daily peak load is 4.8%, and that of the daily valley load is 3.5%, showing good prediction accuracy. The proposed model has reference significance for electric power dispatching.