This chapter summarizes core research advancements in human motion capture and its transformative applications for smart manufacturing. It systematically examines motion capture methodologies encompassing multimodal data acquisition, adaptive pose estimation, and predictive trajectory modeling. These techniques are integrated with teleoperated robots to enable human–robot collaboration. The concept of Phygital Twin is introduced as a paradigm-shifting framework, emphasizing bidirectional knowledge transfer between human operators and their cyber-physical avatars. Future directions prioritize three critical frontiers: (1) Based on human motion digital twin, developing resilient motion capture systems with enhanced environmental adaptability for dynamic industrial settings; (2) Leveraging AI-driven analytics for context-aware robotic responsiveness to human behavioral patterns; (3) Establishing cross-industry scalability through interoperable digital twin architectures, such as human digital twin and physical system digital twin. The discussion advocates for human-centric interface, aiming to accelerate the evolution of cognitive manufacturing ecosystems where human expertise and machine precision achieve symbiotic optimization.