神经网络计算法是模拟人体经络系统活动的机理来研究事物的新方法。本文运用T .Kohonen提出的SOM自组织人工神经网络模型 ,以准噶尔盆地的彩南油田一主力油层为例 ,建立起岩石含油气性评价的人工神经网络模型。实例研究表明 ,人工神经网络法性能良好 ,是一种岩石含油气性评价的有效方法 Neural network algorithm is a new method to study things by simulating the mechanism of meridian system activity. In this paper, using SOM self-organizing artificial neural network model proposed by T. Kohonen, an artificial neural network model for petroleum hydrocarbon evaluation of rock is established by taking a main oil reservoir of Cainan oilfield in Junggar Basin as an example. The case study shows that the performance of artificial neural network is good and it is an effective method to evaluate the hydrocarbon-bearing properties of rocks