The initial and updated Society of Thoracic Surgeons-European Association for Cardiothoracic Surgery (STAT and STAT 2020) and Risk Adjusted Classification for Congenital Heart Surgery-1 and Risk Adjusted Classification for Congenital Heart Surgery-2 scoring systems are validated to predict early postoperative mortality following congenital heart surgery in children; however, their ability to predict long-term mortality has not been examined. We performed a retrospective cohort study using data from the Pediatric Cardiac Care Consortium, a US-based registry of cardiac interventions in 47 participating centres between 1982 and 2011. Patients included in this cohort analysis had select congenital heart surgery representing the spectrum of severity as determined by STAT and Risk Adjusted Classification for Congenital Heart Surgery-1 and were less than 21 years of age. We applied STAT, STAT 2020, Risk Adjusted Classification for Congenital Heart Surgery-1, and Risk Adjusted Classification for Congenital Heart Surgery-2 for prediction of early mortality and long-term postoperative survival probability by surgical risk category. Long-term outcomes were obtained by matching Pediatric Cardiac Care Consortium patients with deaths reported in the National Death Index through 2021. Of 20,753 eligible patients, 18,755 survived the postoperative period and 2,058 deaths occurred over a median follow up of 24.4 years (Interquartile Range: 21–28.4). Each scoring system performed well for predicting early postoperative mortality with the following c-statistics: STAT: 0.7872, Risk Adjusted Classification for Congenital Heart Surgery-1: 0.7872, STAT 2020: 0.7724 and Risk Adjusted Classification for Congenital Heart Surgery-2: 0.7668. The predictive ability for long-term risk of death was as follows: STAT: 0.6995, Risk Adjusted Classification for Congenital Heart Surgery-1 c = 0.6741, Risk Adjusted Classification for Congenital Heart Surgery-2: 0.7156 and STAT 2020: c = 0.7156. Risk-adjusted score systems for congenital heart surgery maintain adequate but diminishing discriminative power to predict long-term mortality. Future efforts are warranted to develop a tool with improved long-term survival prediction.