介绍了根据人工神经网络模型的原理，将地震道信息引入人工神经网络系统进行油气预测的方法。首先从地震时窗内提取出反映储层油气变化的５组共１０多个地震特征参数，然后进行ＢＰ神经网络的自学习，进而结合地质情况和实钻资料进行油气水预测判别，最后将成果绘成图件。将该方法分别应用于四川气区的川南、川中和川东的３个试验区块，均取得与实际或其它预测方法较吻合的结果 The method of introducing seismic trace information into artificial neural network system for oil and gas prediction is introduced based on the principle of artificial neural network model. Firstly, five groups of more than 10 seismic parameters are extracted from the seismic window to reflect the changes of oil and gas in reservoirs. Then BP neural network is used to carry out self-learning, and then the prediction of oil-gas-water water is made based on the geological conditions and real drilling data. Finally, Painted map. Applying this method to the three experimental plots in South Sichuan, Sichuan and Central Sichuan respectively, the results are in good agreement with the actual or other prediction methods