“This study proposes a data-driven resilience planning method for the dynamic characteristics and constraints of distribution networks, constructs a full cycle management system for intelligent devices, develops multi-source fusion elastic reconstruction technology, and forms a closed-loop research from problem diagnosis to strategy implementation. The results show that the dynamic programming model can reduce the load forecasting error by 24%, and the network reconstruction technology can increase the new energy consumption capacity by 18%, providing theoretical support and practical paths for the construction of new power systems.”