The contemporary theoretical framework serves as a core component of political ideology and civic education at the university level institutions. In order to promote the smart advancement of the current theoretical system, this paper constructs BiLSTM-Attention-CRF, a course entity identification model, and BERT-BiLSTM-Attention, an inter-entity relationship extraction model, developed through deep learning, so as to construct the knowledge graph of Civics and Politics courses. Then the course recommendation algorithm KGMO-RS integrating a knowledge graph with multi-objective optimization is proposed, and the type of student satisfaction needs and priority ranking in the Civics classroom are explored by the KANO model. The study indicates that the BiLSTM+Attention-CRF model and the BERT-BiLSTM-Attention model presented in this work surpass competing approaches in the entity identification and inter-entity relation extraction tasks, thereby suggesting that the proposed model is able to construct a higher quality knowledge graph. Meanwhile, the KGMO-RS algorithm outperforms the GCNKG-CR algorithm in terms of HV, IGD and other metrics, which proves that the algorithm demonstrates superior comprehensive search ability and stable convergence behavior throughout the multi-objective optimization process of course recommendation. In addition, drawing on the findings concerning student satisfaction in the Civics classroom, this paper puts forward suggestions to enhance students’ classroom satisfaction in the Civics course across colleges and universities, which provides an approach that can support the intelligent development of the theoretical system in the new era.
