Regional cerebral oxygen saturation (ScO₂) determined by near-infrared spectroscopy, monitoring both arterial and venous blood oxygenation of the brain, could reflect the balance between oxygen delivery and consumption. The aim of this study was to determine the predictabilities of ScO₂ and estimated oxygen extraction ratio (eO₂ER) with outcomes in infants with congenital heart disease (CHD). This study was a two-center, retrospective study of patients at 12 months of age or younger with CHD who underwent cardiac surgery. The primary outcome was a composite of one or more major adverse events (MAEs) after surgery: death from any cause, circulatory collapse that needed cardiopulmonary resuscitation, and requirement for extracorporeal membrane oxygenation. Based on the assumptions of arterial to venous blood ratio, eO₂ER was calculated. A total of 647 cases were included in this study. MAEs occurred in 16 patients (2.5%). There were significant differences in post-bypass ScO₂ [46.61 (40.90, 52.05) vs. 58.52 (51.52, 66.08), p < 0.001] and post-bypass eO₂ER [0.66 (0.60, 0.78) vs. 0.52 (0.43, 0.61), p < 0.001] between patients with MAEs and patients without MAEs. Area under the receiver operating curve (AUROC) of post-bypass ScO₂ was 0.818 (95% confidence interval: 0.747–0.889), AUROC of post-bypass eO₂ER was 0.783 (0.697–0.870) and AUROC of post-bypass maximum serum lactate level was 0.635 (0.525–0.746). Both ScO₂ and eO₂ER, especially after weaning off bypass, are acceptable predictive markers for predicting MAEs after cardiac surgery in infants.