Neonatal mortality remains a huge heath problem in South Africa with leading causes being prematurity, intrapartum hypoxia and neonatal sepsis. The government has made significant progress towards achieving set SDG3 target 3.2 by the end of 2030. However, there is need to reduce absolute numbers of neonatal deaths. Using time series approaches will assist in designing neonatal health policies, decision making and allocation of resources to the maternal and child health program. This research uses annual time series data on neonatal mortality rate (NMR) for South Africa from 1975 to 2019 to predict future trends of NMR over the period 2020 to 2030. Unit root tests have shown that the series under consideration is an I (1) variable. The optimal model based on AIC is the ARIMA (1,1,2) model. The model projections indicate that neonatal mortality will hover around 11 deaths per 1000 live births throughout the forecast period. Therefore, health authorities should craft neonatal policies that will address challenges noted during antenatal, delivery and postnatal periods. There is need to continuously train medical staff on basic and emergency obstetric care & neonatal resuscitation and avail adequate resources at all levels of care especially in primary care health facilities across the country.