“十三五”国家重点研发计划重点专项 (2019YFD0901605)； 辽宁省社会科学规划基金项目（L19BGL016）； 辽宁省自然基金重点项目（20170540005）； 辽宁省教育厅基本科研项目（LQ2017002）
To solve the problem of complex redundancy of water quality early warning indexes,an attribute reduction algorithm based on correlation attribute importance was proposed. On the basis of attribute dependence and attribute importance,the concept of attribute relevance importance was introduced to reduce the decision attribute according to the influence of each conditional attribute on the decision attribute. By calculating the attribute importance and mining the interaction degree among water quality indicators,classification of decision attribute of condition attribute intensity was further distinguished,then the water quality of bass culture was taken as an example to analyze the algorithm,and the UCI data set was used for experimental verification. The results indicate that the algorithm is effective and feasible on redundancy attribute reduction.