By using the daily power load data and meteorological data from June to August from 2010 to 2012, the variations of daily maximum load in muggy weather and sunny hot weather are analyzed. The results show that there were 4 sunny hot processes occurred in June, 5 muggy processes occurred in July and August during the 3 years in Xi’an, and the power load growth was more obvious in muggy processes. Tracking the change of meteorological power load with the daily maximum temperature, when daily maximum temperature reached 33 ℃, the meteorological power load increase rapidly, and the daily maximum temperature of 33 ℃, 35 ℃ and 38 ℃ are three sensitive points of meteorological power load to air temperature variations in Xi’an. Considering the accumulated temperature effect as a forecast factor and using the multiple regression method, a forecasting model for meteorological power load in summer in Xi’an is established. Using the historic data in summer of 2013, the application test shows that the mean relative predicting error of the daily maximum load is 60%。The proposed model can be a good simulation of the actual changes in meteorological power load and provide some guidance for the power supply meteorological service of Xi’an in summer.