In this study, the alpine meadow in Maqin area of Qinghai-Tibet Plateau and the alpine desert steppe in Tuotuo Rivers area were taken as the observation stations. The eddy covariance technique was used to obtain the CO₂ flux, water and energy flux at the level of alpine ecosystem. The data were post-processed by REddyProc and Random Forest (RF), and the variation characteristics of Net Ecosystem Exchange (NEE) at different time scales and different underlying surfaces were analyzed, so as to accurately evaluate and compare the carbon budget under different underlying surfaces. The influence mechanism of meteorological factors on CO₂ flux of alpine meadow and alpine desert grassland was explored, which provided reference and data support for the research on carbon and water cycle of alpine meadow and alpine desert grassland in Qinghai-Tibet Plateau. The results showed that:1) Maqin alpine meadow mainly absorbed carbon sink from June to July, and the absorption peak appeared between 11:00 and 12:00 (Beijing, the same below). However, in March, April, May and August, the main emission was carbon source, and the emission peak appeared between 21:00 and 23:00. The alpine desert of Tuotuo Rivers was mainly absorbed from March to August, showing net carbon sink, and the absorption peak appeared between 13:00 and 14:00; around the whole growing season (March-August), the cumulative NEE of Maqin and Tuotuo Rivers were 79.50 g C/m² and 79.24 g C/m², respectively, showing carbon sink. 2). The correlation between carbon flux and meteorological factors was different in different scales and underlying surfaces. On the hourly scale, the importance of alpine meadow radiation (Rg) to carbon flux was the largest. Evapotranspiration (ET) in alpine desert steppe was the most important factor for carbon flux. On the daily scale, soil water content (Soil _VWC) was the most important factor for carbon flux in alpine meadow, while wind speed (WS) was the most important factor for carbon flux in alpine desert steppe. 3) Ecosystem Respiration (Reco) and Net Ecosystem Exchange (NEE) were significantly restricted by total primary productivity (GPP), and 91% of the GPP contributed to Reco and 9% to NEE; in alpine desert steppe, 40% of the GPP contributed to Reco and 60% to NEE. 4) NEE was significantly correlated with ET, indicating that water condition was the most important factor controlling the changes of carbon and water budget in alpine meadow and alpine desert steppe. 5) Human activities intensified the impact of climate change and the susceptibility of carbon balance, and became the inducing factors of weak carbon sources in March, April, May and August in Maqin.