According to the feature that short term power load is susceptible to meteorological factors, this paper proposes short term power load forecasting model based on similar day and grey theory, firstly tries to get the load data at the same day through the judgment on day type, then deals with the meteorological data sequence by fuzzy clustering analysis, uses the meteorological data at forecasting day and grey correlation method to conduct correlation analysis, chooses the load data which are highly correlated with forecasting day as the load data at similar day, and uses grey forecasting method to forecast short term power load based on similar day load data.Simulation results show that the accuracy of forecasting results is higher after the selection of similar day than without the selection of similar day.