随着减重代谢手术开展数量不断增加，术后体重复胖、代谢性疾病控制不佳及远期并发症等问题日益成为影响病人长期获益的重要因素。复杂减重代谢修正手术是处理上述问题的重要手段，但既往手术造成的解剖结构改变、病人临床情况的个体差异及围手术期风险增加，使术前评估、修正术式选择、手术方案制定和术后管理更为复杂，对临床决策的准确性与规范性提出了更高要求。人工智能能够综合分析病人临床资料、影像学检查、内镜信息及随访数据，为复杂修正手术的个体化决策提供多维度支持，其潜在应用涵盖术前综合评估、围手术期风险预测、修正术式选择、三维辅助规划、术中决策支持及术后随访管理，可贯穿修正手术规划与管理的多个环节。目前，人工智能在医学影像分析和外科风险预测等领域已具备一定研究基础，但直接针对复杂减重代谢修正手术的研究仍较少，相关模型的临床应用价值尚需进一步验证。数据采集、处理及标注缺乏统一标准，多中心研究和临床验证不足，模型可解释性有限，以及医疗责任边界尚未明确，仍是其进入临床实践的主要制约因素。未来应加强减重代谢外科数据库建设，推进临床验证研究，完善人工智能监管体系，并促进人工智能技术与外科医生专业判断有效结合。人工智能作为外科医生的决策辅助工具，有望提升复杂修正手术评估与规划的精准性，推动复杂减重代谢修正手术向更加精准、规范的方向发展。

As the volume of metabolic and bariatric surgery continues to increase, postoperative weight regain, inadequate control of metabolic diseases, and long-term complications have become important factors affecting patients’ long-term benefits. Complex revisional metabolic and bariatric surgery is an important option for addressing these problems. However, anatomical alterations caused by previous procedures, individual differences in patients’ clinical conditions, and increased perioperative risks complicate preoperative assessment, selection of revisional procedures, surgical planning, and postoperative management, thereby placing greater demands on the accuracy and standardization of clinical decision-making. Artificial intelligence (AI) can integrate and analyze clinical data, imaging findings, endoscopic information, and follow-up data to provide multidimensional support for individualized decision-making in complex revisional surgery. Its potential applications encompass comprehensive preoperative assessment, perioperative risk prediction, selection of revisional procedures, three-dimensional assisted planning, intraoperative decision support, and postoperative follow-up management, spanning multiple stages of surgical planning and management. Although AI has an established research foundation in medical image analysis and surgical risk prediction, studies directly addressing complex revisional metabolic and bariatric surgery remain scarce, and the clinical value of relevant models requires further validation. The lack of standardized data collection, processing, and annotation; insufficient multicenter studies and clinical validation; limited model interpretability; and unclear boundaries of medical liability remain major barriers to clinical implementation. Future efforts should strengthen the development of metabolic and bariatric surgery databases, advance clinical validation, improve AI regulatory frameworks, and facilitate the effective integration of AI technologies with surgeons’ professional judgment. As a decision-support tool for surgeons, AI may improve the precision of assessment and planning, and promote the development of complex revisional metabolic and bariatric surgery toward greater precision and standardization.