The reference evapotranspiration (ET ₀) is a crucial indicator for characterizing the potential water consumption of an ecosystem, and the vapor pressure deficit (VPD) serves as a significant measure to assess atmospheric water deficiency. Both indicators are highly correlated with air temperature. However, the response time of ET ₀ to VPD and temperature, as well as the influencing factors, remain unclear. In this study, we employed 21 years of meteorological data from 107 sites in Inner Mongolia, China to analyze the lag response time of ET ₀ to VPD, extreme temperatures (Tmax and Tmin), and vegetation index (NDVI), along with their interactions. The results show that: (1) Between 2000 and 2020, ET ₀, VPD, temperature, and NDVI all exhibited increasing trends; (2) The lagged response time of ET ₀ to VPD was significantly longer than that to temperature (Tmean), with a particularly pronounced effect in high-altitude areas under low-temperature conditions; (3) The potential impact of VPD on ET ₀ was greater than the direct effect of temperature, with ET ₀ responding more to low temperatures than to high temperatures and VPD; (4) An increase in NDVI indirectly extended the lag days of ET ₀ and VPD through extreme temperatures, while the lag days of ET ₀ to extreme temperatures decreased; (5) The lag time of ET ₀ to VPD and temperature was shorter in high-altitude regions, while the lag effect was more prominent in low-altitude areas in the eastern part of the study region. This study reveals the lagged response of ET ₀ to VPD and temperature variations in arid and semi-arid regions, providing new insights into the lagged effects of environmental variables on water and heat fluxes in ecosystems under climate change.