This study demonstrates the potential of machine learning to predict the permeability of soil-fly ash mixtures, thereby promoting fly ash as a sustainable building material. Due to its environmental benefits and enhanced engineering properties when added to mixtures, fly ash, a byproduct of coal combustion, is gaining popularity. Several machine learning algorithms were evaluated, with the linear regression model proving to be the most precise and straightforward. It captured the linear relationship between percentage of fly ash and permeability (RMSE of 6.42 x 10 -06 cm/s and R 2 of 0.811). Training and testing of models utilized a comprehensive database of soil-fly ash mixtures. The implications of the study's findings for engineering and environmental applications are substantial. The model's accuracy in estimating soilfly ash mixture permeability is validated by the excellent correlation between predicted and actual permeability values.