For a toposequence with increasing groundwater influence (Cambisol, Stagnosol, Histosol) and with different groundwater regimes (Histosols 1 and 2) in a forested watershed in the Fichtelgebirge (Germany), the speciation of sulfur (S) and iron (Fe) in the soils was assessed by X-ray absorption near-edge spectroscopy (XANES) after anoxic and conventional oxic sample pretreatments. For samples with anoxic pretreatment, the contribution of reduced inorganic S compounds (monosulfide, pyrite) to total S increased with soil depth for the Cambisol and the Stagnosol, but decreased for the Histosols; the opposite trend was noticed for the contribution of reduced organic S (organic mono- and disulfides, thiols). The contribution of reduced S to the soil S pool increased and the contribution of oxidized S compounds decreased in the sequence Cambisol–Stagnosol–Histosol 1 (permanently anoxic). Histosol 2 (seasonally oxic) showed a markedly larger contribution of oxidized and intermediate S compounds to total S than Histosol 1. The dominating Fe-bearing phases in the Cambisol were Fe(III) oxyhydroxides; the contribution of sulfide-bound Fe was < 5% of total Fe in all horizons. In Histosol 1, the contribution of sulfide-bound Fe increased with soil depth up to 50% in the Cr horizon, whereas in Histosol 2 Fe(III) phases strongly dominated in all horizons. After conventional oxic sample pretreatment, the contribution of reduced inorganic S to total S was markedly decreased in all soils. In the organic surface horizons, the contribution of reduced organic S was increased to the same extent; the contribution of oxidized S (sulfate) remained more or less unchanged. In the mineral soil, the contribution of sulfate and the mean oxidation state of sulfur (MOS) were strongly increased after oxic sample preparation. In Histosol 1, oxic sample pretreatment resulted in oxidation of labile Fe(II) compounds, probably sulfides or Fe(II)–S-org-complexes, to Fe(III). Our study shows that for anoxic wetland soils which contain inorganic sulfide and/or divalent Fe, the exclusion of O2 during the entire period between sampling and analysis is crucial for a correct S and Fe speciation. Only after appropriate sample preparation, clear relationships between the mean oxidation states of S and Fe (MOFe) on one hand and soil hydrological conditions on the other become evident: a concomitant systematic decrease of MOS and MOFe from the well-aerated Cambisol to the permanently anoxic Histosol 1, and larger MOS and MOFe in the seasonally oxic Histosol 2 than in Histosol 1 indicate a close coupling of S and Fe cycling in the soils. Finally, the results of our study suggest that in organic horizons of wetland soils inorganic sulfide S is overestimated and reduced organic S is underestimated by S K-edge XANES, if a significant portion of the thiol groups in reduced organic S is complex-bound to Fe2+ or other chalcophilic metal cations. This is supported by the observation that synthetic organic compounds (cysteine; 1,3,5-trimer-captotriazine [TMT]; ferredoxin) after addition of Fe show spectra with pre-edge peaks at energies < 2472 eV that are typical for inorganic sulfide.