文章针对海洋渔业资源的多种预测算法得到的结果集，提出了以框架的知识表示理论为基础的不确定规则推理预测系统修正算法，利用框架系统ＫＢＩＦ（Ｋｎｏｗｌｅｄｇｅ Ｂａｓｅ ｉｎ Ｆｒａｍｅ）的知识表示理论与黑板结构、不确定推理相结合对预测算法根据不同环境参数进行多种方式的修正，可以对系统预测结果进行有效的分析，以期得到更好的应用。 In this paper, aiming at the result set of many kinds of forecasting algorithms of marine fishery resources, this paper proposes a modified algorithm of uncertain rules reasoning prediction system based on the framework of knowledge representation theory. By using the knowledge representation theory of KBIF (Knowledge Base in Frame) and The combination of blackboard structure and uncertain reasoning can modify the prediction algorithm in various ways according to different environmental parameters, and can effectively analyze the system prediction results in order to get a better application.