Geopolymer concrete (GPC) holds promise as a substitute for traditional concrete, primarily due to its environmental benefits; however, because of the intricate physical and chemical interactions that are a part of the geopolymerization process, developing a reliable forecasting model for compressive strength becomes challenging. Therefore, this study presents a comparative analysis of the neural network model (Artificial Neural Networks, ANN)) and tree-based model (Decision Tree, DT) to forecast the compressive strength properties of ground granulated blast furnace slag (GGBS) based GPC. This study utilized 558 datasets on GGBS-based GPC extracted from various literature sources, incorporating eleven input variables. The predictive performance of all the studied models was validated by analyzing performance metrics such as MSE, MAPE, RMSE, MAE and R2 values. The findings showed that, with a R2 value of 0.943, the ANN model performed better than the DT model in forecasting the compressive strength parameters of GPC made with GGBS. The lower error values (MAE, MSE, MAPE, and RMSE) and higher R2 values strongly indicated the ANN model’s enhanced performance. The results of the input parameter sensitivity carried on the ANN model showed that the specimen’s age, sodium hydroxide molarity, and curing temperature substantially impacted the compressive strength.