为了满足地震解释中无监督型聚类分析的需要，我们提出了一种模糊自组织神经网络的结构和算法，它是在Ｋｏｈｏｎｅｎ自组织网、模糊网及竞争网的基础上发展起来的，它将模糊自组织神经网络中的隶属度概念引入了网络的训练过程，因而收敛速度很快；同时借鉴了竞争网良好的自适应和动态聚类能力及Ｋｏｈｏｎｅｎ网的映射密度图的概念，因此它具有快速处理大容量的数据且分类精度易于控制的特点。此种算法用于地震微相的划分取得良好的效果。 In order to meet the needs of unsupervised cluster analysis in seismic interpretation, we propose a fuzzy self-organizing neural network structure and algorithm based on Kohonen ad hoc network, fuzzy network and competitive network The concept of membership degree in fuzzy self-organizing neural network is introduced into the training process of the network, so the convergence speed is very fast. At the same time, it draws on the concept of self-adaptive and dynamic clustering ability of competitive network and mapping density map of Kohonen network With fast processing of large-capacity data and the classification accuracy is easy to control features. This algorithm is used to obtain good results in the separation of seismic microfacies.