在智能决策支持系统（ＤＳＳ）研究中，知识表示、自然语言理解以及相应的推理系统是急待解决的几大问题。传统的精确推理已难以适应计算机应用发展的需要。近年来在推理机的设计中发展了规则与结论都不唯一的非精确、非单调推理。但如何测定结论的可信度、可用度仍是个问题。本文探讨基于信息论引入测度的概念，以解决非精确推理结论的可信度问题。 In DSS research, knowledge representation, natural language understanding and corresponding reasoning systems are several major issues to be solved urgently. The traditional precise reasoning has been difficult to meet the needs of the development of computer applications. In recent years, the design of inference engine has developed inaccurate and non-monotonic reasoning that is not unique in rules and conclusions. But how to measure the credibility of the conclusion, the availability is still a problem. This article explores the concept of introducing measure based on information theory to solve the problem of credibility of inexact reasoning conclusion.