The investigation of the fractionation of S compounds in forest soils is a powerful tool for interpreting S dynamics and S biogeochemistry in forest ecosystems. Beech stands on high pH (nutrient-rich) sites on Flysch and on low pH (nutrient-poor) sites on Molasse were selected for testing the influence of stemflow, which represents a high input of water and dissolved elements to the soil, on spatial patterns of sulfur (S) fractions. Soil cores were taken at six distances from a beech stem per site at 55 cm uphill and at 27, 55, 100, 150 and 300 cm downhill from the stem. The cores were divided into the mineral soil horizons 0–3, 3–10, 10–20, 20–30 and 30–50 cm. Soil samples were characterized for pH, Corg, pedogenic Al and Fe oxides and S fractions. Sequential extraction by NH4Cl, NH4H2PO4 and HCl yielded readily available sulfate-S (RAS), adsorbed sulfate-S (AS) and HCl-soluble sulfate-S (HCS). Organic sulfur (OS) was estimated as the difference between total sulfur (ToS) and inorganic sulfur (RAS + AS + HCS). Organic sulfur was further divided into ester sulfate-S (ES, HI-reduction) and carbon bonded sulfur (CS). On Flysch, RAS represented 3–6%, AS 2–12%, HCS 0–8% and OS 81–95% of ToS. On Molasse, RAS amounted 1–6%, AS 1–60%, HCS 0–8% and OS 37–95% of ToS. Spatial S distribution patterns with respect to the distance from the tree stem base could be clearly observed at all investigated sites. The presented data is a contribution to current reports on negative input–output S budgets of forest watersheds, suggesting that mineralization of OS on nutrient rich soils and desorption of historic AS on nutrient-poor soils are the dominant S sources, which have to be considered in future modeling of sulfur.