Aiming at the main point of the traditional teaching mode in which the language ability cultivation and the Civic and Political education are separated, this study proposes a CD-CAT system for college English based on the blended teaching mode. It constructs a course Civic and Political objective system containing a three-dimensional dynamic objective matrix, and develops a dual-coordinate question bank system integrating linguistic knowledge points and Civic and Political elements. Intelligent algorithms such as Shannon entropy and KL information quantity are introduced to optimize the topic selection strategy, and a dynamic learning path planning model is established by combining Q learning. Under the condition of item parameter U(0.05,0.4), the cognitive model including basic knowledge, basic skills and cognitive process dimensions with a total of 22 attributes is constructed. Student A is selected for application case study. Overall, Student A has a better mastery of basic knowledge and basic skills, with 7 attributes fully mastered, but a poorer mastery of cognitive process attributes, with no attributes fully mastered and more than 50% of attributes in a state of no mastery.
