The growing silver digital divide, resulting in the exclusion of elderly populations from electronic government (e-government) services, has become a pressing issue in aging societies worldwide. While many existing efforts focus on making individual technologies more age-friendly, they often overlook deeper challenges, such as how to dynamically adapt services, protect user privacy, and coordinate across different domains. To tackle these issues, in this study, we introduce an elderly-centered adaptive governance framework that integrates the Internet of Things (IoT) with multi-agent systems (MASs). The proposed framework is built on a four-layer architecture. The perception layer uses low-intrusive wearable devices and environmental sensors to collect multimodal data. The network layer combines 5G edge computing with lightweight blockchain (IOTA Tangle) to reduce latency and ensure data trustworthiness. At the cognitive level, digital twins and federated learning enable distributed, privacy-preserving decision-making. Finally, the application layer uses context-aware interfaces to adapt interaction logic in real time. This approach brings together several innovations. It introduces a closed-loop perception–decision–response model, showcasing how IoT and MASs can work together to support proactive service delivery. It also includes mechanisms for adaptive user interfaces, multimodal biometric authentication, and the integration of digital and physical public services tailored to the diverse needs of older adults. Furthermore, the framework incorporates smart contract-based accountability (chain of responsibility) and a policy pathway that addresses algorithm transparency and regional access inequality. With a scalable and inclusive approach to digital governance, this framework helps move e-government services beyond basic functionality toward more responsive, human-centered systems. It also provides valuable insights for the development of smart cities and more inclusive digital societies.