Background: Critical illness is collection of conditions and disorders that have impact on the occurrence of organ dysfunction, prolonged morbidity and mortality. Score of mortality objectively calculates the severity of disease and predicts risk of mortality according to the clinical condition of patient. Today, no standardized scoring system that used in PICU for initial assessment of patients in Indonesia yet. Objective: Comparing mortality?s predictor ability between PRISM IV and PELOD 2 score as predictors of mortality in critically ill children at Haji Adam Malik General Hospital Medan. Subject and Methods: Prospective cohort study was conducted in PICU of Haji Adam Malik General Hospital from September 2020 to March 2021. Characteristics of the sample were obtained through the medical record of patients treated in PICU. Assessment of PRISM IV and PELOD 2 score was carried out on the first and third days of patient being admitted in PICU. Results: Of the 35 children who were treated, 15 (42.9%) children were treated with respiratory problems. There were 18 (51.4%) children died. It was found that there was no significant difference mean score of PRISM IV and PELOD 2 score on first and third days of examination (p>0.05). It was also found that PRISM IV could predict mortality of children with critical illness on the first and third days (AUC 49.7%; 57.2%). PELOD 2 could predict mortality on first and third days (AUC 65.4%; 69.3%). Conclusion: PRISM IV on the first and third days has lower accuracy than PELOD 2 in predicting mortality in critically ill children at Haji Adam Malik General Hospital Medan.