Abstract Background Magnesium Sulphate (MgSO4) has become a standard practice for fetal neuroprotection prior to delivery of very-preterm babies due to an association with MgSO4 and reduction of cerebral palsy at 2 years of age in randomized clinical trials. Objectives In a prospective cohort of very preterm babies, we aimed to define the association between antenatal MgSO4 administration and neurodevelopmental outcome. Design/Methods We included a prospective cohort of 265 infants born ≤ 32 weeks’ gestation recruited at three tertiary neonatal intensive care units (NICU). Infants underwent two brain MRIs, one in the initial weeks after birth and one at term-equivalent age. Neurodevelopmental follow-up assessments were completed using the Bayley-3 at 3 years. Univariate analysis was undertaken to determine whether MgSO4 was associated with outcome. Multiple maternal and neonatal factors were analyzed according to their relationship with outcome. Gestational age was accounted for in all models. Maternal factors associated with MgSO4 administration were hypertension, multiple births and antenatal corticosteroid administration. Neonatal factors were sex, neonatal resuscitation, major surgery, and moderate-severe white matter injury (volume >40mm3). Multivariable linear regression models of maternal and neonatal risk factors were used to account for potential confounders based on those factors that were significantly (p<0.05) associated in univariate analysis. Results On univariate analysis, MgSO4 administration was inversely associated with 36-month motor scores (p<0.05). On the multivariable analysis, no maternal confounders had an impact on the relationship between MgSO4 and outcome (Table 2). MgS04 was no longer statistically related to any Bayley outcomes after adjusting for confounders at 36 months (Table 2). Neonatal surgery was related to poorer 36-month motor and cognitive outcomes. Conclusion Antenatal MgSO4 exposure was not associated with improved measures on the Bayley scales at 3 years after adjusting for potential confounders. However, neonatal interventions remained significant risk factors for poorer neurodevelopmental outcome.