According to a survey by the World Health Organization, up to 75% of the population is in a sub healthy state. Long term sub-health status not only increases the risk of chronic diseases, but also leads to low personal health management efficiency and continuous depletion of social medical resources. The fundamental reason lies in the dual constraints of users’ own cognitive deficiencies in health management and the passivity and fragmentation of traditional health management models. With the deep integration of intelligent technology and health management, digital health management has gradually become a new path to solve the sub-health dilemma. The global digital health market has formed a mature industrial ecosystem. However, existing systems generally face development bottlenecks such as shallow interaction and rigid data, which exposes the structural deficiencies of traditional systems in cognitive guidance and emotional interaction. It is urgent to build new methods that can achieve intelligent data transformation and humanized interaction. This project adopts a complete R&D process of “requirement research functional design technical implementation testing verification”: firstly, conducting requirement research and sorting through multiple methods; Secondly, based on the results of the requirements analysis, design the system’s functional architecture and interactive interface; Then complete the AI Agent interface docking, localized model deployment, and front-end and back-end system development; Finally, closed-loop verification is completed through user experience testing and system availability evaluation. At the same time, a complete architecture covering software, hardware, and backend services has been constructed.