Soil suction is the major property that controls the behaviour of unsaturated soils. Suction estimation is challenging both in the lab and the field. Various instruments to measure the suction have been developed with the recent technological advancements. Nonetheless, there are still limitations in regards to the reliability, cost, suction range, accessibility, scope of activity, and appropriateness for use either within the field or lab settings. The filter paper method is probably the simplest procedure to measure the suction for the entire range both in the field and the lab. Nevertheless, the procedure takes time. To alleviate the requirement for conducting this test, it becomes imperative to develop mathematical predictive models for soil suction. In this study, a detailed survey was carried out across the Free State province, South Africa, and sampling points identified. Samples were tested for their geotechnical properties. The influence of the geotechnical properties on soil suction was studied. Multivariate regression analysis was carried out utilizing MINITAB 18 program to develop the mathematical predictive models.