Forecasting core inflation in South Africa is important for economic planning. Researchers used different models to predict core inflation, finding that small time-varying parameter vector autoregressive models work best. Models with changing error patterns perform better than those with consistent errors. Keeping the forgetting factor constant throughout the forecast period is more effective than allowing it to change over time. Including information on economic growth and interest rates helps accurately forecast core inflation. However, the relationship between these variables needs to be modeled in a time-varying way.