In the evolving landscape of 5G Heterogeneous Cloud Radio Access Networks (HC-RAN), efficient Device-to-Device (D2D) communication is paramount. This paper presents the Enhanced Adaptive Algorithm (EAA), a novel energy-efficient solution tailored for multi-hop D2D communication in 5G HC-RAN. EAA addresses the dual challenges of maintaining high energy efficiency and robust network performance under diverse conditions. The algorithm’s core innovation lies in its dynamic interference management and balanced computational strategy. It leverages machine learning for predictive interference mitigation and employs a hybrid computational model, distributing tasks between edge devices and central servers. This approach ensures efficiency without compromising scalability. EAA’s adaptability is further enhanced by its sophisticated use of Channel State Information (CSI), incorporating real-time updates and a robust design tolerant of CSI inaccuracies. It also features scenarioaware optimization and AI-based analysis of user and traffic patterns, making it highly responsive to varying network environments. Incorporating multi-objective optimization, EAA balances energy efficiency with key network performance metrics, employing Pareto optimization techniques to navigate complex trade-offs. Its modular design and continuous learning component future-proof the algorithm, enabling easy integration with emerging technologies. Extensive simulations validate EAA’s effectiveness, showcasing notable improvements in energy efficiency and overall network performance for multi-hop D2D communication in HC-RAN. This work not only advances 5G network optimization but also lays the groundwork for future enhancements in wireless communication systems.