Wetlands store large amounts of carbon stocks and are essential in both global carbon cycling and regional ecosystem services. Understanding the dynamics of wetland carbon exchange is crucial for assessing carbon budgets and predicting their future evolution. Although many studies have been conducted on the effects of climate change on the ecosystem carbon cycle, little is known regarding carbon emissions from the alpine wetlands in arid northwest China. In this study, we used an automatic chamber system (LI-8100A) to measure ecosystem respiration (ER) in the Bayinbuluk alpine wetland in northwest China. The ER showed a significant bimodal diurnal variation, with peak values appearing at 16:30 and 23:30 (Beijing time, UTC + 8). A clear seasonal pattern in ER was observed, with the highest value (19.38 mu mol m(-2) s(-1)) occurring in August and the lowest value (0.11 mu mol m(-2) s(-1)) occurring in late December. The annual ER in 2018 was 678 g C m(-2) and respiration during the non-growing season accounted for 13% of the annual sum. Nonlinear regression revealed that soil temperature at 5 cm depth and soil water content (SWC) were the main factors controlling the seasonal variation in ER. The diurnal variation in ER was mainly controlled by air temperature and solar radiation. Higher temperature sensitivity (Q(10)) occurred under conditions of lower soil temperatures and medium SWC (25% <= SWC <= 40%). The present study deepens our understanding of CO2 emissions in alpine wetland ecosystems and helps evaluate the carbon budget in alpine wetlands in arid regions.