Digital transformation and industrial upgrading have put forward new requirements for the precise and personalized training of secondary vocational network technology talents. To address the current problems in secondary vocational computer network course teaching, such as significant differences in students’ basic foundations, insufficient adaptability of teaching resources, and rigid learning paths, this study constructs an AI-driven personalized teaching model centered on learner profiles. By collecting multi-source teaching data, the model dynamically builds fine-grained learner profiles. On this premise, it focuses on exploring personalized teaching path generation methods based on reinforcement learning and knowledge graphs, as well as adaptive resource recommendation mechanisms integrating content correlation, collaborative filtering, and sequence patterns. The research aims to form a teaching closed loop of “data perception—profile portrayal—intelligent decision-making—precision intervention”, providing a data-intelligence empowered practical path for the teaching reform of secondary vocational specialized courses, so as to improve teaching efficiency, stimulate students’ potential, and promote the cultivation of high-quality technical and skilled talents.