Objective: Small for gestational age (SGA) fetuses are at increased risk of admission to the neonatal unit, even at term. We aimed to develop and validate a predictive model for the risk of prolonged neonatal unit admission in suspected SGA fetuses at term. Methods: A single-center cohort study of singleton pregnancies with SGA fetus, defined as estimated fetal weight (EFW) less than the 10th centile, at term. The variables included known risk factors for neonatal unit admissions: maternal characteristics, EFW, abdominal circumference (AC), fetal Dopplers, gestational age (GA) at delivery, and intrapartum risk factors (meconium, pyrexia). Logistic regression analysis was used for model building and the prediction models were validated internally using bootstrapping. Results: Seven hundred and one SGA pregnancies at term were included; 5.9% had prolonged neonatal unit admission (>48 h). The multivariable model (AUC 0.71; 95% CI: 0.63-0.79) included GA at delivery <39 weeks (OR 2.76; 95% CI 1.23-6.04, p = .011), cerebroplacental ratio (CPR) multiples of median (MoM) (OR 0.21; 95% CI 0.05-0.79, p = .023), and EFW below the third centile (OR 2.43; 95% CI 1.26-4.68, p < .007). The combined model showed a sensitivity 30.9% (95% CI: 16.6-45.2%) for a fixed 10% false positive rate. Conclusion: The prediction model shows good accuracy and good calibration for assessing the risk of neonatal unit admission in suspected SGA fetuses. It has the potential to be used for patient counseling, determining the timing of delivery and the individual risk.Brief rationale Objective: The objective of this study is to determine the factors associated with prolonged neonatal unit admissions in small for gestational age fetuses at term. What is already known: Fetal weight and Doppler parameters are associated with adverse outcome in small for gestational age fetuses. However, most studies use composite outcome criteria by combining neonatal unit admission with adverse delivery outcomes. A comprehensive model combining antenatal and intrapartum variables is also lacking. What this study adds: Our model describes the association of antenatal and intrapartum variables with prolonged neonatal unit admission without using a composite adverse outcome measure. Estimated fetal weight, gestational age at delivery, and the cerebroplacental ratio can be used to estimate the risk of prolonged neonatal unit admission. The risk estimation can be useful for patient counseling and to determine the time of delivery.