Organic inputs were used for 10 years on a French vineyard topsoil to improve structural stability and thus to protect against erosion. The three types of organic inputs (mulches) included: conifer compost, CC (100 m3 ha−1 every 3 years); conifer bark, CB (300 m3 ha−1 every 5 years); and cereal straw, S (10 t ha−1 every 2 years). The other two types of organic inputs were cover crops of clover (C) and fescue (F). The impacts of these organic inputs on soil organic carbon (SOC) content, wettability (capillary rise and X-ray photoelectron spectroscopy (XPS)) and structural stability were studied. The SOC content was twice as large in the CC, C and F topsoils (SOC content of 2.56–3.24%) as in the reference (R) topsoil (SOC content of 1.39%). Both apparent contact angle (θ) and surface OH:C mass ratio indicated that the R and S topsoils were hydrophilic (θ of 27.4–33.4°, surface OH:C ratio of 3.20–4.41), whereas the CB, C and F topsoils were partially hydrophobic (θ of 69.1–79.8°, surface OH:C ratio of 1.36–2.86), and the CC topsoil had intermediate values (θ of 46.9°, surface OH:C ratio of 2.43–2.81). Moreover, the greater the θ value, the smaller the water sorptivity value and the greater the proportion of water-stable aggregates, Agw. The increase in SOC content had beneficial effects on Agw, particularly for the partially hydrophobic C and F topsoils (Agw of 22.3–44.5%) against the hydrophilic R and S topsoils (Agw of 8.2–12.7%). Development of hydrophobicity, correlated with the decrease in the surface OH:C ratio and the increase in the C–O, C–N proportion on surface C, should be attributed to humified organic matter or/and to plant and microbial polysaccharides. As the XPS and aggregate stability data describe soil physical processes at small scales (nm to mm), we suggest an experimental and modelling framework for upscaling these results for practical improvement and management of vineyard soils.