||The control of soil moisture, vegetation type, and prior land use on soil health parameters of perennial grass cropping systems on marginal lands is not well known. A fallow wetness-prone marginal site in New York (USA) was converted to perennial grass bioenergy feedstock production. Quadruplicate treatments were fallow control, reed canarygrass (Phalaris arundinaceae L. Bellevue) with nitrogen (N) fertilizer (75 kg N ha−1), switchgrass (Panicum virgatum L. Shawnee), and switchgrass with N fertilizer (75 kg N ha−1). Based on periodic soil water measurements, permanent sampling locations were assigned to various wetness groups. Surface (0–15 cm) soil organic carbon (SOC), active carbon, wet aggregate stability, pH, total nitrogen (TN), root biomass, and harvested aboveground biomass were measured annually (2011–2014). Multi-year decreases in SOC, wet aggregate stability, and pH followed plowing in 2011. For all years, wettest soils had the greatest SOC and active carbon, while driest soils had the greatest wet aggregate stability and lowest pH. In 2014, wettest soils had significantly (p < 0.0001) greater SOC and TN than drier soils, and fallow soils had 14 to 20% greater SOC than soils of reed canarygrass + N, switchgrass, and switchgrass + N. Crop type and N fertilization did not result in significant differences in SOC, active carbon, or wet aggregate stability. Cumulative 3-year aboveground biomass yields of driest switchgrass + N soils (18.8 Mg ha−1) were 121% greater than the three wettest switchgrass (no N) treatments. Overall, soil moisture status must be accounted for when assessing soil dynamics during feedstock establishment.