The potential for human exposure to arsenic (As), a group A carcinogen, has increased tremendously because of the encroachment of suburban areas into former agricultural lands, where arsenical pesticides were used extensively prior to the 1990s. Many baseline risk assessments of As-enriched sites assume that all (100%) As present in the soil is bioavailable because of the high expenses and logistical difficulties associated with conducting site-specific in-vivo studies to measure bioavailable As fractions. This assumption seriously overestimates the actual risk because various geochemical forms of As are insoluble in human gastric/intestinal juices and are not likely to be bioavailable. A laboratory incubation study was conducted to estimate bioavailable As as a function of soil properties. Three different soil types were chosen based on their potential differences with respect to As reactivity: an acid sand with minimal As retention capacity, a sandy loam with relatively high concentration of Fe/Al oxides, and an organic (muck) soil with high Fe/Al, Ca/Mg content with potential for higher As retention capacity. Our results show that soil properties had a marked impact on geochemical speciation of As. Hysteritic adsorption of As onto Fe/Al oxides significantly decreased bioavailability of As in both Millhopper and Pahokee Muck soils, which have high concentrations of Fe and Al. High concentration of organic matter in Pahokee Muck soil solubilized As, resulting in higher bioavailability in Pahokee Muck soil compared to the Millhopper soil, although Pahokee Muck soil has higher concentrations of Fe and Al. The soils varied significantly in terms of soil speciation and bioavailability of As, indicating significant effect of soil properties on As biogeochemistry.