Soil samples (n = 45) and water samples (n = 111) were collected in the coastal zone of SW Bangladesh in wet (November) and dry (May) seasons in 2016 to identify the factors influencing soil arsenic concentrations. Soils are entisols formed from recently deposited, predominantly silt-sized sediments with low carbon concentrations typical of the local mangrove forests. Arsenic concentrations in bulk soil are higher in November than in May and vary little between sites. Arsenic concentrations in deionized H2O extracts are ~2 orders of magnitude lower, indicating only ~1% of As is soluble. Water samples show that As concentrations are highest in groundwater from tubewells. Bulk soil As is positively correlated with As concentration in irrigation water, suggesting that As from irrigation water is added to the soil. Unlike other water types, As in rice paddy water is much higher in the wet season, consistent with some fields being irrigated with tubewell water. Arsenic in rice paddy water increases by soil sulfide dissolution, and decreases by dilution during the monsoon. Water soluble As in rice paddy soils is positively correlated with S and DOC concentrations in rice paddy soil extracts due to sulfide dissolution and complexation with DOC. Thus, waterlogging of rice paddy soils leads to reducing conditions, the absence of ferric oxyhydroxides that could sorb As, and the presence of sulfides that incorporate As. As soil pH increases from 7 to 8, KD(soil/extract) increases, consistent with the observed positive correlation between irrigation water pH and bulk soil As concentration. Arsenic bioavailability could be decreased through soil aeration (draining and tilling) and by avoiding the use of groundwater for rice paddy irrigation.