Over 90% of Canadian peatlands are in the boreal region, which is especially susceptible to climate warming, particularly during the winter or non-growing season (NGS). It has been reported that peatlands act as a source of carbon dioxide (CO2) during the NGS and how these emissions compare to the growing season (GS) CO2 uptake remains uncertain. In addition, measuring CO2 fluxes during the NGS is challenging due to harsh weather conditions in boreal regions, limiting the availability of eddy covariance data during this period. To combat these challenges, remote sensing and modeling may be a valuable method to estimate the NGS CO2 fluxes. In this study, we defined the start of the NGS as the first day of three consecutive days with temperatures below 1°C. We selected peatland sites across Canada and acquired seven years (2015-2021) of derived CO2 data from NASA's Soil Moisture Active Passive Carbon Net Ecosystem Exchange (SMAP-NEE) dataset. However, SMAP-NEE tended to overestimate in-situ NEE. Therefore, we adjusted SMAP-NEE using a linear regression with in-situ NEE, which reduced the overestimation. The adjusted dataset showed differences in SMAP-NEE between sites, years, and seasons (p<0.001). Moreover, our data analyses showed that NGS CO2 emissions represent a variable proportion of the GS CO2 uptake, and when these NGS emissions were accounted for, the annual CO2 sink strength was reduced. Our findings highlight the importance of considering the NGS when constructing annual budget estimates. However, we stress the importance of further validation of the SMAP-NEE algorithm in northern peatlands.