Objective To establish the basic principles for constructing, validating, and implementing clinical intervention of a digital twin model for spinal rehabilitation, and to provide references for the development and application of related systems. Methods Forty-two experts from fields such as spinal surgery, neurosurgery, rehabilitation medicine, pain medicine, medical imaging, nursing, medical engineering and artificial intelligence, medical ethics, and hospital management participated in the consensus formulation. The working group retrieved relevant literature published from January 2014 to May 2026, and conducted two rounds of inquiries after discussing the topics. In the first round, 42 experts were sent questionnaires, and 40 valid responses were obtained; in the second round, 40 questionnaires were sent, and 39 valid responses were received, with response rates of 95.2% and 97.5% respectively. Using a 5-level Likert scale, items with scores of 4?5 were considered as“ agree”, and items with an agreement rate of 80% or higher were included in the consensus. Results A total of 16 consensus opinions were formed, covering the definition and application scope of the digital twin, system architecture, data collection and governance, model validation and update, rehabilitation intervention, clinical application, and quality management. The agreement rates of each item in the first round ranged from 82.5% to 100%, and increased to 89.7% to 100% in the second round. Conclusion Spinal rehabilitation digital twin should not only be used for data display; its core lies in continuously updating the model based on specific patients and incorporating the predicted or scenario-driven results into the rehabilitation process reviewed by doctors and therapists.