Aims: To investigate the pharmacokinetics and safety of prolonged paracetamol use (>72 h) for neonatal pain. Methods: Neonates were included if they received paracetamol orally or intravenously for pain treatment. A total of 126 samples were collected. Alanine aminotransferase and bilirubin were measured as surrogate liver safety markers. Paracetamol and metabolites were measured in plasma. Pharmacokinetic parameters for the parent compound were estimated with a nonlinear mixed-effects model. Results: Forty-eight neonates were enrolled (38 received paracetamol for >72 h). Median gestational age was 38 weeks (range 25–42), and bodyweight at inclusion was 2954 g (range 713–4750). Neonates received 16 doses (range 4–55) over 4.1 days (range 1–13.8). The median (range) dose was 10.1 mg/kg (2.9–20.3). The median oxidative metabolite concentration was 14.6 μmol/L (range 0.12–113.5) and measurable >30 h after dose. There was no significant difference (P >.05) between alanine aminotransferase and bilirubin measures at <72 h or >72 h of paracetamol treatment or the start and end of the study. Volume of distribution and paracetamol clearance for a 2.81-kg neonate were 2.99 L (% residual standard error = 8, 95% confidence interval 2.44–3.55) and 0.497 L/h (% residual standard error = 7, 95% confidence interval 0.425–0.570), respectively. Median steady-state concentration from the parent model was 50.3 μmol/L (range 30.6–92.5), and the half-life was 3.55 h (range 2.41–5.65). Conclusion: Our study did not provide evidence of paracetamol-induced liver injury nor changes in metabolism in prolonged paracetamol administration in neonates.