The research on optimization of teaching resource push based on adaptive learning path explored the implementation method of personalized learning in the intelligent education environment. The study proposed a learning path feature framework with "how to learn" and "what to learn" as the core, and designed a resource recommendation and knowledge point matching model by combining static and dynamic features. In terms of "how to learn", the accuracy of recommendation is improved through the diversification of multimedia resources and learning preference analysis; in terms of "what to learn", dynamic feature analysis supports real-time adjustment of learning content to achieve the optimal matching of knowledge points. In addition, the study also explored the application of digital technology in the lifelong education system to provide flexible and efficient learning solutions for individuals at different learning stages. Finally, this study constructed a teaching resource push model that integrates resource push, path optimization and learning behavior feedback, providing theoretical and technical support for personalized education practice.