Peatland restoration seeks to re-establish a number of ecosystem services, with carbon (C) sequestration key following the inclusion of emissions from rewetted peatlands in national GHG reporting. Few studies address the effects of restoration on GHG fluxes or quantify annual emissions from peatlands undergoing restoration. This research focuses on the response of CO2 and CH4 fluxes from degraded lowland peatlands in Wales to restoration measures; vegetation removal, topsoil removal and rewetting. Fluxes were quantified from three locations using closed-chamber measurements, with annual fluxes estimated using empirical models. The effect of vegetation removal on CO2 and CH4 fluxes was closely coupled to water table depth, with the method of vegetation removal determining plant community composition. The removal of nutrient-enriched topsoil resulted in a net C emission, but this was assumed to be temporary as a small net C sink was observed where vegetation recolonisation occurred. The effect of rewetting was studied at a cutover raised bog and a grassland on peat. CO2 and CH4 fluxes from microforms created during historical hand-cutting were analogous to natural hummock-hollow complexes; CO2 fluxes were greater from higher drier areas, whereas lower wetter conditions favoured higher CH4 emissions. Rewetting at the cutover bog increased CH4 fluxes where the water table was close to the surface and aerenchymatous vegetation was present, however under inundated conditions where vegetation was absent, CH4 fluxes were relatively low. The rewetted grassland on peat was a significantly stronger CO2 sink compared to a shallow-drained grassland, attributed to high nocturnal CO2 emissions from the latter. CH4 emissions were greater from the rewetted grassland, though the high CO2 uptake resulted in a net GHG balance. Annual emissions derived from this research are comparable to UK lowland peatlands, but lower than IPCC emission factors, highlighting the importance of deriving UK-specific emission factors for future reporting.