<p>Expansive soils, known for their swelling characteristics, are common in civil engineering projects worldwide. These soils can lead to issues such as differential settlement, ground heaving, pavement cracking, and canal liner damage. Therefore, evaluating the swelling potential (swelling percentage) and swelling pressure is essential in foundation analysis. However, laboratory tests used to determine these parameters are both costly and time-consuming, often requiring complex equipment, high computational expenses, and skilled personnel. Developing reliable prediction models that relate swelling properties to rapid, cost-effective, and non-destructive tests could greatly enhance structural design practices. In this study, four multivariable mathematical models of multivariable linear Developing regression (LR), pure quadratic (PQ), interaction (IA), and full quadratic (FQ) were utilized, incorporating liquid limit, plasticity index, dry density, and water content as input parameters for approximately 400 data points from literature to estimate swelling pressure and swelling potential. The results indicated that all the employed models provided estimation with R² values greater than 0.7; however, the FQ model outperforms the other models with an R² value of 0.8 for both swelling parameters. Finally, the sensitivity analyses were conducted. The results indicated that the liquid limit is the most crucial parameter for swelling pressure, while the swelling potential was mainly affected by moisture content. These findings fill a significant research gap and provide a practical, efficient alternative to conventional testing methods. The developed models offer significant potential for enhancing foundation design and infrastructure safety by enabling faster and more cost-effective assessment of expansive soils.</p>