Introduction. Prolonged mechanical ventilation (PMV) is a well-recognized factor as a quality metric for pediatric cardiac surgical programs. Most of the risk factors for PMV are described and analyzed. Some authors had established predictive models to detect proactively patients in risk for PMV. This study aims to develop a new predictive model, based on vasoactive-ventilation-renal (VVR) score, for PMV after congenital heart surgery (CHS) in pediatric patients. Material and Methods. Medical fi les of patients 0-18 y who underwent heart surgery in 2016 and 2017 were reviewed. Patients that met the inclusion criteria were studied. PMV was defi ned as invasive mechanical ventilation ≥ 96 h. The patients were divided in two groups according to duration of mechanical ventilation: group 1-patients with PMV, group 2-patients without PMV. The focus was set on VVR score and fl uid overload in the fi rst 48 hours after the operation. Data were presented as medians with IQR or as means ± standard deviation. A non-parametric Mann-Whitney U test, binary logistic regression test and ROC curve analysis integrated in the statistical software SPSS 24.0 were used. A value of P < 0.05 was considered signifi cant. Results. 438 patients were operated in 2016 and 2017 and 384 of them were included in the study. 80 patients (20.8%) needed PMV (group 1) and 304 (79.2%) did not need PMV (group 2). There was a statistical signifi cance between group 1 and group 2 concerning the peak VVR for the day of operation 58,25(33,48) vs. 25,65(19,8) and cumulative fl uid overload in % for the fi rst 48hours +2,54(13,29) vs. – 1,19(3,4). After combining this two factors in a predictive model, the ROC curve analysis showed AUC 0,903 (95% CI 0,863-0,944) with sensitivity of 86.25% and specifi city of 82,57%. Conclusion. Combining VVR and cumulative fl uid overload, resulted in establishment of a new reliable predictive model for PMV after CHS in pediatric patients in our Center.