In contemporary manufacturing, the integration of human and robot capabilities has become imperative for enhancing efficiency and flexibility. This paper presents a novel approach to human–robot collaboration within intelligent assembly systems, leveraging the synergies of digital twin technology and knowledge graphs. By employing digital twins to simulate and optimize assembly processes, coupled with knowledge graphs to represent and manage domain-specific information, our proposed framework enables seamless interaction and collaboration between human operators and robotic agents. Through real-time data exchange and decision-making support, the system facilitates adaptive task allocation, error detection, and process optimization, ultimately leading to improved productivity and product quality in assembly operations.