In order to overcome the drawbacks of traditional forecasting methods, a method of POL consumption forecasting based on weighted grey relational analysis and fuzzy case-based reasoning combined retrieval is proposed. Firstly, a case retrieval model based on weighted grey relational analysis is constructed, and the weight of grey relational coefficient is determined by using information entropy theory. Then, a case retrieval model based on fuzzy set theory is constructed, and the weight of feature attributes is determined by improved analytic hierarchy process. Finally, based on the above two retrieval models, a combined retrieval model is constructed by using grey relational analysis method, and the POL consumption is forecasted based on the retrieval results. A numerical example is given to demonstrate high accuracy of the retrieval method and the feasibility and practicability of the prediction method.