As a component of macroeconomic assumptions, economic growth is used in the state budget drafting. However, the nominal GDP data used to calculate the economic growth value has a release lag in the first week of the second month after the quarter ends. This research aims to find the best nowcasting model for MIDAS regression, RFR, LightGBM, SVR, MLP, and ensemble methods. Then, based on the best model, the nominal GDP value in the 2023 fourth quarter and the 2024 first quarter are predicted. Overall, the MLP model with variable selection using SHAP values has the best evaluation indicators, therefore this model is used to predict the nominal GDP in both quarters. Using the MLP model, the predicted value of nominal GDP in both quarters is quite accurate when compared to nominal GDP data that have been released by BPS-Statistics Indonesia.