在我国进行路网规划过程中，由于在采用路网决策支持系统时展线系数难以确定，从某种程度上影响了路网的决策。本文从展线系数的特性分析入手，首先采用模糊聚类分析法进行地形分类，然后建立多元线性回归模型来确定各个区域的展线系数，有利于为路网决策提供更为准确的信息。

In the railway network planning, effective information for the project decision cannot be provided because it is hard to get the expanding coefficient. This paper first assort the topography of different district with the fuzzy classification, and then the expanding coefficient is gotten by the model of linear regression, so we can get more accurate information for the network planning.