Extensive arid and semi-arid ecosystems in Central Asia are threatened by aridification and desertification owing to intensified evaporation and extreme climates. To understand the mechanisms of water vapor (H2O) transformations and assess the water balance under climate change in Central Asian grasslands, knowledge of H2O ecosystem-scale flux and its seasonal and interannual dynamics is important. Based on the eddy covariance technique, this study measured the five-year H2O flux over the Bayinbuluk Grassland in Central Asia and investigated its environmental controls. The results showed that the grassland was a net source of H2O flux, emitting 1432 +/- 93 mm y(-1) from 2017 to 2021, with a mean annual precipitation of 237 +/- 69 mm. Seasonal changes in H2O fluxes were higher during the growing season than that during the non-growing season, with annual maxima generally occurring from July to August. A clear unimodal diurnal pattern in the H2O flux was observed during both seasons from 2018 to 2021, with peak values appearing at approximately 14:30. We further conducted a wavelet analysis on this long-term quasi-continuous H2O flux time series and investigated its temporal variability and wavelet coherence with environmental variables. Daily periodicity in H2O fluxes was detected during most of the growing season. The variations in H2O fluxes were in phase with changes in air temperature and solar radiation on a daily timescale, with relative humidity showing a negative correlation with H2O flux. Changes in precipitation, air temperature, soil temperature, and photosynthetically active radiation exhibited stronger positive correlations with H2O fluxes at monthly and annual timescales than daily timescales. In addition, annual evapotranspiration increased at a rate of similar to 100 mm y(-1) during the study period, despite precipitation and air temperature showing no apparent increasing trends. Grassland ecosystems in arid Central Asia are expected to emit more H2O under a warming climate, leading to greater water scarcity and heightened aridity. Our study highlights the importance of conducting long-term continuous eddy covariance and time-series analyses to enhance our understanding of the temporal variability in grassland H2O exchanges.