Lei Wang, Yina Jiang, Jialong Zhu, Kunyang Lin, Xiaofei Ma, Jun He, Bingheng Lu
Digital Twin (DT) provides a pivotal solution to space manufacturing bottlenecks through high-fidelity simulation and closed-loop control. This paper systematically reviews DT for space manufacturing. It first clarifies the unique conceptual framework and challenges arising from microgravity, resource limitations, and high autonomy requirements. Next, it traces DT’s evolution – from industrial static prototypes, through virtual-physical interaction exploration, to near-real-time closed-loop systems – and analyzes the migration of industrial technologies to space and nascent on-orbit applications. Critically, the study focuses on four technical pillars: high-fidelity multi-physics modelling under microgravity, near-real-time simulation and interaction, multi-source sensing and sparse data fusion under resource constraints and deeply-integrated autonomous collaborative decision-making. Applications in space additive manufacturing, on-orbit assembly, maintenance, repair, and in-situ resource utilisation (ISRU) demonstrate DT’s core value in driving the ‘design-manufacturing’ closed-loop. Concurrently, this review further identifies key bottlenecks, including inadequate model fidelity in extreme environments, sparse sensing data, limited on-board computing power, Earth-space communication latency, and high systems integration/verification costs. Finally, intelligent enhancement pathways are proposed – including PIML-empowered AI-driven dynamic modelling, Earth-space collaborative computing, digital thread standardisation, and human-cyber-physical collaborative decision-making – to advance space manufacturing from ground-dependent verification toward on-orbit autonomous closed-loop operations, laying the foundation for sustainable deep-space exploration.