The integration of advanced technologies into manufacturing processes is critical for addressing the complexities of modern industrial environments. In particular, the realm of human–robot interaction (HRI) faces the challenge of ensuring that human operators can effectively collaborate with increasingly sophisticated robotic systems. Traditional interfaces often fall short of providing the intuitive, real-time interaction necessary for optimal performance and safety. To address this issue, we introduce a novel system that combines digital twin (DT) technology with augmented reality (AR) to enhance HRI in manufacturing settings. The proposed AR-based DT system creates a dynamic virtual model of robot operations, offering an immersive interface that overlays crucial information onto the user’s field of vision. This approach aims to bridge the gap between human operators and robotic systems, improving spatial awareness, task guidance, and decision-making processes. Our system is designed to operate at three distinct levels of DT functionality: the virtual twin for in-situ monitoring, the hybrid twin for intuitive interaction, and the cognitive twin for optimized operation. By leveraging these levels, the system provides a comprehensive solution that ranges from basic visualization to advanced predictive analytics. The effectiveness of the AR-based DT system is demonstrated through a human-centric user study conducted in manufacturing scenarios. The results show a significant reduction in operational time and errors, alongside an enhancement of the overall user experience. These findings confirm the potential of our system to transform HRI by providing a safer, more efficient, and more adaptable manufacturing environment. Our research contributes to the advancement of smart manufacturing by evidencing the synergistic benefits of integrating DT and AR into HRI.