Climate conditions affect energy consumption for heating in areas that experience harsh winters. Skillful prediction of energy demand ensures a reliable energy supply and protects vulnerable populations from cold weather and reduces waste. Here, we develop a statistical model for prediction of the winter seasonal electricity consumption over the UK using a multiple linear regression technique. The model uses the autumn conditions of Arctic sea-ice concentration, stratospheric circulation, and sea-surface temperature, skillful predictors for the North Atlantic Oscillation (NAO) reported in a previous study. The model skillfully predicts the electricity consumption one month in advance, with statistically significant anomaly correlation between the predicted and the observed energy consumption time series. Further analysis showed that the energy consumption in the UK is significantly correlated to surface air temperature, dew point depression, and windspeed, which are all related to the NAO and can be skillfully predicted using the same predictors. Additionally, we found skillful prediction of surface air temperature, dew point depression, and windspeed over larger areas where the NAO is influential, which implies the predictability of electricity consumption in these regions. The statistical model demonstrates the usefulness of climate information in energy management and has profound socioeconomic implications.