Fen peatlands represent a globally important carbon (C) store, while also providing highly productive agricultural land. Drainage of these organic soils is required to create conditions suitable for crop growth, but this results in substantial greenhouse gas (GHG) emissions. One potential GHG mitigation option is to raise the groundwater table to reduce the duration and volume of peat exposure to aerobic conditions. However, the trade-off between maintaining food production and securing ecosystem function under a high water table (WT) presents a serious challenge for both land managers and policy makers. Therefore, we conducted a controlled mesocosm experiment to investigate the effects of WT elevation (from −50 cm to −30 cm) under three contrasting scenarios: (i) WT raised throughout the year, (ii) WT raised in the winter only, and (iii) WT raised in the growing season only. We measured GHG emissions, nitrate, ammonium and dissolved organic C concentrations in soil solution, alongside the yield of a commercially important crop (lettuce). Raising the WT throughout the year reduced lettuce yields by 37% and reduced CO2 emissions by 36% without changing the loss rates of N2O or CH4. Raising the WT only in the winter did not significantly reduce crop yield, but still suppressed CO2 emissions during the fallow period (by 30%). Raising the WT only in the growing season reduced root growth and CO2 emissions (by 27%), but had no major effect on lettuce yield. In conclusion, the present study shows that raising the groundwater table in the non-growing season reduced GHG emissions without negatively affecting lettuce yields, and may therefore represent a viable GHG mitigation option for agricultural peatlands.