Intelligent Text Mining Based Financial Risk Early Warning System With increasing risks faced by financial firms and markets, text mining is playing more and more important role. However, because of the complexity between a text and the consequences caused by the text, it is hard for traditional text mining techniques to meet the needs. In order to solve this problem, we focus on the employment of intelligent agents to enhance the capability of text analysis. In our system, various classes of intelligent agents are proposed for financial risk early warning. The text mining agent gives influence measurement, the relationship agent generates relationship measurement between the text and the target, and the reliability agent measures whether the text is trustable or not. Collaboration within such agents is able to produce more effective and accurate early warning messages.