This paper takes the results of the practice of intelligent construction of intelligent Civics teaching mode in colleges and universities as the theoretical basis and constructs the evaluation system of the effectiveness of intelligent Civics teaching based on the dynamic clustering-gray correlation entropy method. The original data are standardized by the method of polar deviation to determine the association coefficient of the evaluation subject, and according to the clustering process, the clustering order table is listed to realize the construction of the association matrix. F-statistics determine the optimal threshold to prevent human subjective factors from affecting the clustering results. The rough set theory is used as a mathematical tool for data analysis and processing, and the screening of indicators is completed through the approximation of indicators, and the corresponding index system for the evaluation of wisdom and politics is established. The index system is applied to the actual evaluation of the quality of intelligent civics teaching to analyze its practical effects. The results show that the proportion of all the indicators in the evaluation index system of intelligent Civics teaching constructed in this paper that have been evaluated as more important or above ranges from 80% to 100%. Civics teachers’ mining of the elements of Civics, the extension of teaching content, and teaching content delivery are improved by 54.96%, 48.39%, and 45.07%, respectively, after the transformation of the intelligentized mode, and the quality of teaching has been significantly improved.