Aiming at the problems of complex content and abstract concept of the Computer Network course, and difficulty for students to establish overall cognition, this paper proposes a teaching framework that integrates knowledge graph and visualization technology. Firstly, through ontology modeling and multi-source data extraction, a course knowledge graph covering 180 core concepts and multi-level semantic relationships was constructed. Secondly, hierarchical, interactive, and thermal views were developed to dynamically present the knowledge structure, protocol process, and learning data. Finally, the visual graph was embedded into the flipped classroom and experimental links to realize personalized learning path recommendations and real-time mastery feedback. Teaching practice shows that the framework significantly improves students’ overall cognitive level and learning motivation, and provides a reproducible and generalizable method framework and practical evidence for the smart teaching of computer courses.