Due to massive weather sensitive load in terminal equipment, the features of load changing with meteorological factors become more and more significant. Based on the virtual dummy regression forecasting model which has low complexity and strong ability to explain, the factors of human power consuming behavior, temperature and relative humidity and their across effects are introduced to the model. The dynamic mapping relationship between load and these factors is deeply mined. In terms of human power consuming behavior, dummy variables are introduced to describe the load changing rule in year, week and day. Load specificity of holidays is considered and quantified to improve the practicability of the model. In terms of temperature, cubic function model is used to describe the relationship of load and temperature, and the recency effect of temperature is put forward that means the lag hourly temperature information is introduced to the model. In terms of relative humidity, a new modeling method is put forward with quadratic function model of variables. At last, the proposed model is validated by actual historical data from 2006 to 2010 in a certain area. The prediction results show that the proposed model can reflect the load changing trend in various date types, the method is suitable for municipal power grids with little change in meteorological factors.