Abstract: The numerous existing methods for prediction of oil and gas production are reviewed, and the principle of conventional combination forecast method is explicated. The method for determining optimal weights in combination forecast based on least square method is improved. A novel method based on BP artifcial neural networks is proposed, which is also applied to the case study in this paper. The results show that this novel method or model a8 a preferential technique can be used to effectively improve the accuracy for prediction of oil-gas production.

Key words: production, prediction, accuracy, neural network, model