Experiments conducted to determine swelling characteristics are costly and time-consuming. Therefore, recently, researchers have been using artificial intelligence methods to calculate the relationship between some easily measurable properties of soils and swelling characteristics. This study used data from studies on expansive soils reported in the literature; the relationship between swelling pressure and the plasticity index, initial dry unit weight, and initial soil water content was examined with ML-based methods. For this purpose, Random Forest (RF) and Gradient Boosting (GB) algorithms were preferred. In the RF algorithm, R2 and RMSE values for the training and testing stages were calculated as 0.995, 0.940, and 4.843, 12.122, respectively. For the GB algorithm, these values were obtained as 1.000, 0.974, and 3.395, 8.034, respectively. The fact that R2 values are very close to 1 and the errors are quite small shows that both algorithms successfully estimate the swelling pressure. However, based on these statistical indicators, it can be said that the model using the GB algorithm is more efficient.