PURPOSE: With increasing use of extracorporeal membrane oxygenation (ECMO) for acute respiratory distress syndrome (ARDS) in children, development of standardized strategies for survival prediction has become crucial; however, this has not been accomplished yet. We evaluated the adult scoring systems for survival prediction used for their applicability in pediatric ARDS and validated them. METHODS: A total of 11 children with ARDS receiving ECMO from 2013 to 2014 were evaluated with adult scoring systems, including the Respiratory Extracorporeal-membrane-oxygenation Survival Prediction (RESP), the PRedicting dEath for SEvere ARDS on VV-ECMO (PRESERVE), and the ECMOnet scores. We compared the scores on these scales and the clinical characteristics between survivors and nonsurvivors. RESULTS: Eight of the 11 children died (72.7%). The PRESERVE score (survivors vs. nonsurvivors: 2 vs. 5.25, P=0.048), and the ECMOnet score (4.1 vs. 5.63, P=0.048) were lower in survivors than in nonsurvivors. They correctly predicted mortality prediction. There was no significant difference in the RESP score between survivors and non-survivors (−4.33 vs. −2.62, P=0.63). The parameters that showed significant differences in this study were peak inspiratory pressure, platelet, and delta neutrophil index. All children who were under immunocompromised conditions, such as those with tumors, or underwent hematopoietic stem cell transplantation died. The immunocompromised status should be considered an important factor for survival prediction in children with ARDS. CONCLUSION: This is the first pilot study to apply the survival prediction scoring system to pediatric ARDS with ECMO. It is necessary to establish and modify the survival prediction score system for pediatric ARDS with ECMO.