This study aims to investigate the coordination‐behavior patterns of control crews in digital nuclear power plants (NPPs) during emergencies from a network perspective. We observed and coded 12 coordination processes (each from one crew) of handling simulated emergencies on a full‐scope dynamic simulator in Tianwan NPP of China. By calculating the proportion of coordination breakdowns and referring to the subjective evaluation of instructors, these control crews (all male) were classified into two performance levels (high and low). To compare the coordination‐behavior patterns between high‐ and low‐performing crews, we conducted social network analysis based on the number and direction of coordination behaviors. By examining intracrew relations in team coordination, this study indicated that high‐performing crews exhibited higher cohesion and more balanced behavioral patterns in team coordination than low‐performing ones. By investigating coordination‐behavior patterns from the hierarchical structure of relations in the team, this study revealed that high‐performing crews exhibited higher degree of team autonomy and self‐management among junior operators, as well as reactor operator (RO)‐centered pattern, whereas the low‐performing crews exhibited senior reactor operator‐centered pattern. For researchers, the results advance the understanding of the team coordination mechanism in NPP control rooms during emergencies and enrich team coordination theory in process control industries from the network perspective. In practice, this study suggested that the team coordination training of NPP control crews might be centered around junior operators (especially RO), with focuses on facilitating information sharing and mutual assistance between junior operators and enhancing the awareness of active cooperation of crew members. The results of the study provide nuclear instructors with practical reference to team coordination training and processes optimization.