The invention provides an electricity consumption prediction method and device. The method comprises the steps of obtaining real-time data based on variables obtained by pre-training; predicting by adopting a Ridge regression method based on the real-time data corresponding to the variable to obtain an electricity consumption prediction value, wherein variables obtained through training are variables with the highest relevancy with electricity consumption obtained through two-step variable screening on the basis of historical electricity consumption data, economic data, traffic data and meteorological data, the prediction effect is good, and the prediction precision is improved; when the technical scheme provided by the invention is applied to long-term prediction, the fitting effect on asequence with relatively large random volatility is relatively good, the method is suitable for occasions with large data volume and occasions with small data volume, over-fitting on a training set isavoided in the occasions with small data volume, the fitting effect is relatively good, and the prediction precision is greatly improved. According to the technical scheme provided by the invention,the influence of random factors is small, and the reliability of prediction results is high.