The Adaptive Renewable Energy Management System (AREMS) is a new framework that has been proposed to solve the problems of integrating renewable energy in urban power systems. Through the use of dynamic load optimisation, wireless power transfer and more reliable features, AREMS improves the overall energy utilization, reliability of power grids and resistance to disturbances. The analytic system that is proposed in this work uses the Long Short Term Memory (LSTM) networks for predictive models to keep track of the energy demand and supply, and the Pelican Optimization algorithm (POA) to optimize these opera-tions. Simulation results demonstrate significant achievements: an estimated 15% increase in energy efficiency, 25% decrease in System Average Interruption Du-ration Index (SAIDI) and 30% reduction in System Average Interruption Fre-quency Index (SAIFI). Integrating WPT achieved transmission efficiency of up to 92% within a transmission range of one meter, which is viable for rising urban applications. Moreover, AREMS achieved 98% accuracy in fault detection, and it has exceeded a traditional system as a predictive monitor and achieving grid sta-bility. These outcomes confirm the use of the presented system as a scalable ap-proach to integrate dynamic load management, renewable energy integration, and real-time fault detection. AREMS provides best practice guidelines for adaptive energy management, allowing the creation of new, powerful, and sustainable ur-ban energy systems.