This paper proposes a new approach to improve short-run energy demands forecasting using temperature-price cross response function, which exploits the fact that the relative price of energy has different effects on the energy demands at different temperatures. Using this new approach, we provide empirical analyses of Korean short-run electricity and city gas demand forecasting. The estimation and forecasting results suggest that there should be a significant impact from the relative price on the energy demand especially for heating demand. Our forecasting procedure is shown to perform significantly better than other methods that are commonly used for short-run energy demand forecasts in the literature.