传统的岩性识别技术主要基于统计学理论，如贝叶斯方法、回归方法等，近年来人工神经网络方法如反向传播算法（ Ｂａｃｋ － Ｐｒｏｐａｇａｔｉｏｎ ， Ｂ－ Ｐ） 也应用于岩性识别，取得了一定的效果。用 Ｋｏｈｏｎｅｎ 提出的自组织特征映射神经网络对测井数据进行岩性识别，该方法具有较强的自组织性、自适应性，有较高的容错能力。与 Ｂ－ Ｐ 算法相比较，计算量小，收效速度快，且不需要已知的先验信息而自动确定分类类别。结果表明与统计方法、岩性录井分析结果一致。 Traditional lithology identification techniques are mainly based on statistical theory, such as Bayesian method and regression method. In recent years, artificial neural network methods such as Back Propagation (B - P) are also applied to lithology identification. A certain effect. Using the self-organizing feature map neural network proposed by Kohonen for lithology identification of well logging data, this method has strong self-organization, adaptability and high fault tolerance. Compared with the B-P algorithm, the computational load is small, the efficiency is fast, and the classification category is determined automatically without known prior information. The result shows that it is consistent with the statistical method and lithological well logging analysis.