This document describes research using an artificial neural network to predict Indonesia's stock exchange composite index using macroeconomic variables. The research used historical data for exchange rates, interest rates, inflation rates, and money supply as inputs to train an ANN model. The best-performing model used a time delay of two months and achieved a prediction accuracy of 96.38% with an average error of 0.0045. The researchers concluded the ANN successfully predicted index movements and that macroeconomic indicators are suitable for this type of prediction. They suggest adding other variables or using daily data for more complex and beneficial predictions.