Bumetanide is a loop diuretic that is proposed to possess a beneficial effect on disorders of the central nervous system, including neonatal seizures. Therefore, prediction of unbound bumetanide concentrations in the brain is relevant from a pharmacological prospective. A physiologically‐based pharmacokinetic (PBPK) model was developed for the prediction of bumetanide disposition in plasma and brain in adult and paediatric populations. A compound file was built for bumetanide integrating physicochemical data and in vitro data. Bumetanide concentration profiles were simulated in both plasma and brain using the Simcyp PBPK model. Simulations of plasma bumetanide concentrations were compared against plasma levels published in the literature. The model performance was verified with data from adult studies before predictions in the paediatric population were undertaken. The adult and paediatric intravenous models predicted pharmacokinetic factors, namely area under the concentration–time curve, maximum concentration in plasma and time to maximum plasma concentration, within two‐fold of observed values. However, predictions of plasma concentrations within the neonatal intravenous model did not produce a good fit with the observed values. The PBPK approach used in this study produced reasonable predictions of plasma concentrations of bumetanide, except in the critically ill neonatal population. This PBPK model requires more information regarding metabolic intrinsic clearance and transport parameters prior to further validation of drug disposition predictions in the neonatal population. Given the lack of information surrounding certain parameters in this special population, the model is not appropriately robust to support the recommendation of a suitable dose of bumetanide for use as an adjunct antiepileptic in neonates.