Abstract: Project risk prediction is a very important issue，project managers very concern about it.But there is not a method accepted by most of the scholar until today.This paper tries to use AntClust algorithm to resolve this problem and through a numerical example to explain the whole process，the result of the indicating is good.It finds out that compared with SVM the AntClust algorithm has the characteristics of simple operation.

Key words: AntClust, Support Vector Machine（SVM）, risk forecast

摘要： 工程项目风险预测是一个非常重要的问题，项目经理都很关心，但是至今仍然没有一种项目风险预测方法得到学者们公认。尝试将蚁群化学聚类算法用于这一问题，并通过一个算例来说明整个过程，结果显示预测效果良好。针对此问题将蚁群化学聚类算法与SVM进行了比较，表明了与SVM相比蚁群化学聚类算法具有简单易操作的特性。

关键词: 蚁群化学聚类, 支持向量机, 风险预测