This study uses annual time series data on under five mortality rate (U5MR) for Vietnam from 1964 to 2020 to predict future trends of U5MR over the period 2021 to 2030. Residuals and forecast evaluation criteria indicate that the applied ANN (12, 12, 1) model is stable in forecasting under five mortality. The multilayer perceptron neural network projections revealed thatU5MR will hover around 22 deaths per 1000 live births throughout the out of sample period. Therefore, health authorities in Vietnam must address all the challenges that hinder the successful implementation of SDG3 with the aim of improving child survival.