Background: Colorectal cancer (CRC) is a leading cause of cancer-related morbidity worldwide. Surgical resection is the main curative treatment, but postoperative complications such as anastomotic leak, surgical site infection, and mortality remain common. Traditional risk prediction tools have limited accuracy. Artificial intelligence (AI), including machine learning (ML) and deep learning (DL), may improve prediction of postoperative outcomes following CRC surgery. Methods: A systematic review of studies published up to December 2025 was conducted using PubMed, Embase, Scopus, and the Cochrane Library. Eligible studies applied ML or DL models to real-world patient data and reported predictive performance metrics, including area under the curve (AUC). Results: Thirty studies were included. AI-based models consistently outperformed conventional risk assessment tools, with reported AUC values ranging from 0.80 to 0.87. Common ML techniques included XGBoost, random forest, and artificial neural networks, which showed good performance in predicting anastomotic leak, surgical site infection, reoperation, and mortality. Intraoperative AI applications, such as indocyanine green fluorescence imaging, were associated with lower anastomotic leak rates. Models incorporating natural language processing of electronic health records improved early infection detection. Most studies were retrospective and single-center. Conclusion: AI models demonstrate strong potential for predicting postoperative complications after CRC surgery but require multicenter prospective validation before routine clinical use. References: Mohamedahmed AY, et al. Int J Surg. 2025;111:8550–8562. Tian Y, et al. BMC Med Inform Decis Mak. 2024;24:11