Abstract: The long-standing problems of product quality and production efficiency in additive manufacturing have persistently hindered its industrial-scale development. Digital twin technology offers a new paradigm for process optimization and quality control through virtual-physical interaction and data-driven approaches. This paper aimed to summarize the implementation methods of digital twins in additive manufacturing, the conceptual evolution, core elements, and research progress were systematically reviewed. Firstly, the synergistic potential between the characteristics of additive manufacturing processes and digital twin technology was elucidated. Secondly, from the two core elements of models (geometric, physical, behavioral, and rule-based) and data, the key pathways for constructing digital twin systems were analyzed. Finally, focusing on the four core functionalities of monitoring, prediction, decision-making, and control, the breakthroughs and limitations of existing technologies in multi-scale simulation, real-time monitoring, and intelligent regulation were summarized, and future development directions were proposed.