Abstract:   In order to make a scientific energy strategic decision, it is vitally important to accurately forecast the energy supply and demands. In this paper, time series methods are applied to forecast the supply and demands of refined oil for a province in China. The study is conducted by using data of petroleum supply and demands collected in the Province during the last 10 years. With the data, statistical error is analyzed and related statistics test is conducted by using Crystal Ball prediction software. Then, four time series methods are applied to forecast the demands for the next five years. By comparison, it shows that the result obtained by using double exponential smoothing is the best. The result is meaningful for the sustainable development of the Province.

Key words:  supply and demands of petroleum, time series analysis, forecast