The study of Civic Education in Colleges and Universities in the Age of Intelligence is a rational reflection on the integration of intelligent technology and education in a contemporary society characterized by informationization, digitization, and intelligence. This paper improves the traditional YOLOv5 algorithm and integrates the CA attention mechanism into the YOLOv5 algorithm to construct Civic and Political Intelligent Education based on classroom state recognition. Subsequently, a systematic test of the A-YOLOv5 system for the smart classroom was carried out, comparing the traditional YOLOv5 algorithm, the unimproved Mobilenet- YOLOv5 algorithm, and the A-YOLOv5 algorithm, which was analyzed and concluded to be superior to the other two algorithms in terms of the speed of detection and the accuracy of recognition. At the same time, the classroom effect of A-YOLOv5 was tested, and it was concluded that the A-YOLOv5 model could accurately reflect the status of classroom students. In this paper, the A-YOLOv5 algorithm improves the recognition rate of the state of listening, playing cell phone, raising hand, writing, and sleeping in the classroom of Civics and Political Science teaching compared with the YOLOv5 algorithm and evaluates the recognition accuracy of 85.42%. The model of Civics education under big data can improve the system’s recognition of students’ classroom behavior and concentration, which facilitates the teaching mode of Civics education and makes it easy for teachers to change their teaching methods in response to students’ behavior.