Abstract: According to the characteristics of power and oil consumption, an energy consumption model based on the least square support vector machine was established, using the flowrate as the independent variable. In the model, RBF kernel function was chosen. The results of this method showed that the prediction accuracy for oil consumption was lower than power consumption, and the reason was that the influencing factors of oil consumption were more complicated. Posteriori estimation was used to test the prediction accuracy of power and oil consumption by this model, and the results were both "good". An application example showed that the model was simple, and the prediction result was precise. The method can be used to predict the power and oil consumption of crude oil pipeline.