Based on the data ofdaily demand for electricity from June toAugustduring the period of2005- 2007 in Shijiazhuang, the meteorological electric quantity variedwithweather factorswas setapart, and the correlation coefficients betweenmeteorologicalelectric quantity on rainy days or cleardays and keymeteorological factors on the same periodwere calculated respectively, especially themeteorological electric quantity changing ruleswith temperature, humidity and precipitationwere studied. The results show that themeteorological electric quantity correlated positivelywith temperature, butnegativelywith relative humidity on rainy days. The correlation coef-
ficients between daily change of electric quantity and daily variation ofmeteorological elementswere calculated also, and itwas found that they had good correlation. Based on statistical analysis, themultiple regression predictionmodels forweather electric quantity and electric quantity daily change under the influence of compositiveweather elements on rainy days and cleardays respectivelywere established by usingOrigin 7. 5 software. The regression statistics and VAR analysis both indicate that all equations have passed the F test
withα=0. 0005. Themultiple correlation is higher than single correlation, and the consistent rate is highwhichmeans thismodel can be used as a reference for electric power department to attemper the electric power reasonably.