Alpine wetlands play a sensitive function in global carbon cycle during the ongoing climate warming, yet the temporal patterns of carbon dynamics from in situ ground-based long-term observations remains unclear. Here, we analyzed the continuous net ecosystem CO2 exchange (NEE) measured with the eddy covariance technique over an alpine peatland on the northeastern Qinghai-Tibetan Plateau from 2007 to 2016. The wetland acted as a net CO2 source with a positive NEE (120.4 ± 34.8 gCm-2year-1, Mean ± S.D.), with the mean annual gross primary productivity (GPP) of 500.3 ± 59.4 gCm-2year-1 and annual ecosystem respiration (RES) of 620.7 ± 74.2 gCm-2year-1. At the seasonal scale, the classification and regression trees (CART) analysis showed that aggregated growing season degree days (GDD) was the predominant determinant on variations in monthly NEE and monthly GPP. Variations in monthly RES were determined by soil temperature (Ts). Furthermore, non-growing season Ts had a significant positive correlation with the following year annual GPP (p<0.05). Non-growing season RES only accounted for about 25% of annual RES, but had significant correlation with annual RES and annual NEE (p<0.05). The further partial correlation analysis showed that non-growing season air temperature (Ta, p = 0.05), rather than precipitation (PPT, p = 0.25) was a predominant determinant on variations in annual NEE. Our results highlighted the importance in carbon dynamics of climate fluctuations and CO2 emission from the non-growing season in alpine wetlands. We speculated that the vast peadlands would positively feedback to climate change on the Tibetan plateau where the non-growing season warming was significant.