中文摘要：

为了提高汇流模拟的精度，在对洪水分类及汇流模型参数优选方法研究的基础上，先采用模糊聚类分析对各历史洪水进行分组，然后在各组内采用优选算法分别优选汇流参数。其优势在于分类方法原理简单，计算量小，但能减少不确定因素的影响，突现分类后各组资料的相似性，最终能够达到提高模拟精度和优选效率的目的。结合棠荆流域水文模拟实例验证了效果。

英文摘要：

In order to increase the precision of the confluent simulation, after studying the methods of flood classification and the confluent parameter＇s optimum-seeking, this article uses the fuzzy cluster analysis to divide historical floods into several groups, and then optimizes the confluent parameters respectively in each group. The superiority of this method is not only the simple principle of taxonomic approach but also the small computation load, by which can reduce the influence of uncertainty , highlight the similarity of materials and improve the precision of the analogy and the optimum-seeking efficiency finally. The effects will be verified by the example of jingtang basin＇s hydrologic analogy.