We studied the restoration success of a cut-away peatland 10 years after restoration by comparing the vegetation and CO2 dynamics with those of a pristine peatland of similar nutrition level and climate. Vegetation and CO2 dynamics were monitored during one growing season. We used DCA (detrended correspondence analysis) and diversity indices to study the vegetation composition within and between the sites, and non-linear regression models to estimate the seasonal CO2 fluxes and balances of the sites. Based on both DCA and diversity indices, the study plots in the restored site differed more in the vegetation composition than the study plots in the pristine site. The variation in the CO2 fluxes and balance was greater in the restored than in the pristine site, resulting from the heterogeneous vegetation in the restored site. The seasonal net CO2 balance was positive (sink) at both sites, the restored site binding on average 500 ± 410 g CO2/m2 and the pristine site 390 ± 265 g CO2/m2 (statistically not different, p = 0.575). The results indicate that the restoration of the vegetation composition is still incomplete but the vegetation coverage is sufficient for the restored site to function as a sink of atmospheric CO2.