Abstract: The research investigation in the hydrology data mining consists of six procedures：object understanding，data preparation，data pretreatment，model creation，valuation and interpretation，and knowledge application.The adoption of public data processing software and the algorithm mining software has succeeded in connecting each software with the data，forming the loose coupling system of the hydrologic data mining.In the whole procedure of the hydrologic data mining，the data mining of data processing mode is applied to the hydrology domain，as well as adopting some of the data processing techniques in the hydrology domain，resulting in the combination of the data mining domain with the profession domain.The application research is based on the hydrologic data of two different representative regions from the national hydrologic basic data base in Jiangsu Province with a strict control according to the data mining processes，starting from the search of the hydrologic similar year，and thus the data mining research in the hydrology domain is carried out.After an all-direction analysis and comparison of the hydrologic similar year，with the method of the data mining，and the cluster analysis for the condensation algorithm of the hierarchical cluster，the discovery of the result agrees with the hydrology domain expert knowledge.

摘要： 研究探索了包括目标理解、准备数据、数据预处理、建立模型、评估解释、知识应用等水文数据挖掘的六个过程。并采用公共的数据处理和挖掘算法，实现各过程之间数据无缝连接，形成了松散耦合的水文数据挖掘系统体系框架。在实施水文数据挖掘过程中，将数据挖掘的一些数据处理方式应用到了水文领域，同时也采用了水文领域中的一些数据处理技术，实现了数据挖掘领域与专业领域的数据处理和评价方式融合。通过选取江苏省国家水文数据库中的两个不同代表性区域的水文资料，严格按照水文数据挖掘的过程控制，以水文相似年查找为突破口，实施数据挖掘。在全方位地对结果分析、对比和评价后发现，以数据挖掘的方法，采用聚类分析中分层聚类的凝聚算法，进行水文相似年查找所发现的结果与水文领域专家知识基本相符。