证据冲突是多证据融合中各种合成方法的主要制约问题,本文引入K-L信息距离函数描述证据间的冲突特征,通过构造距离矩阵定义独立证据在整个系统的冲突程度,实现对证据冲突的系统化识别.模拟实验表明:K-L信息距离的有效冲突识别,能完善D-S理论合成规则的应用约束,对正常冲突证据合成得到优化收敛结果,对高度冲突隔离证据的特征分析提取价值信息. Conflict of evidence is the main constraint of various synthetic methods in multi-evidence fusion. In this paper, KL information distance function is introduced to describe the characteristics of conflict between evidences. By constructing distance matrices, the degree of conflict of independent evidences in the whole system is defined and the evidence conflict is systematized The simulation results show that the effective conflict identification of KL information distance can improve the application constraints of DS theory synthesis rules, optimize the convergence results of normal conflict evidence synthesis, and extract the value information from the feature analysis of highly conflicting evidence of isolation.