Currently, economic new normal and frequent extreme weather make influencing factors of power demand more diversified. Forever, integrated forecast methods of such two power demand indicators as power consumption and max power load arenot scientifically formed. Main influencing factors of power demandfrom economic and meteorological dimensions are systematicallyanalyzed and a forecast model system of provincial grid’s annualpower demand is established. The instance of Anhui province demonstrates that the key of improving forecast accuracy is to identify critical influencing factors regardless of basing forecast model on multiple linear regression or intelligent algorithm. Mean forecast error rates of power consumption and max power load remain are both within 2% in this model, becoming an effective method for electricity market analysts to forecast annual power demand.