Fluctuations of land water availability (LWA, defined as the difference between precipitation and evapotranspiration) may increase the drought risks and widely affect management of water resources. Globally, LWA are mainly produced by the model simulations due to rarely evapotranspiration observation available. Based on AMIP simulations at four horizontal resolutions (i.e., 1.425°, 1.00°, 0.50°, and ~0.25°) for 1982-2014, this study mainly investigates the impact of resolution on the LWA variability (sLWA, represented by the standard derivation of monthly LWA anomalies) and its hydrological components over global land areas. Results indicate that all four simulations generally reproduce the spatial patterns and zonal-mean structures of sLWA observed in the reference, while higher resolution has lower biases. The simulation at 0.25° is found to perform best, especially over the Sahel, North Africa, and middle-high latitudes of Eurasia. Decomposition of the terrestrial water budget components reveal that refined model resolution strengthens the runoff–LWA covariability, while weakens soil moisture–LWA covariability. In contrast, low-resolution simulation presents too strongly negative soil moisture–LWA coupling and overestimates the effects of soil moisture memory, especially over arid regimes. The combined hydrological effect associated with model resolution contributes the sLWA over global land areas. These findings underscore the importance of high-resolution ESM for accurately capturing terrestrial water resources related to the regional land-atmosphere coupling.