A comprehensive study of expansive soil behavior includes understanding the surface phenomena of clay particles within the soil matrix. This research studies four micro-scale properties of four remolded expansive soils––matric suction, pH, surface conductance and percentage of montmorillonite––in order to predict soil swelling. An approach to approximate surface conductance is presented. Linear regression analyses were undertaken in an attempt to predict percent swell and swell pressure based on each of these micro-scale properties. Matric suction was found to be the most accurate predictor of the swelling behavior of the studied soils, which were initially compacted at optimum moisture content. Surface conductance, which is a combination of specific surface area, cation exchange capacity and cation mobility, also gave good predictions, except for one soil with high acidity.