Plants exhibit varying strategies for optimizing the trade-off between CO2 uptake and water loss through transpiration in response to increasing air or soil dryness. Anisohydric plants generally keep their stomata open to maintain or enhance carbon uptake, but this exposes them to a greater risk of hydraulic failure. In contrast, isohydric plants tend to maintain hydraulic integrity by enforcing stricter stomatal and xylem regulation, albeit at the expense of reduced carbon assimilation. Information on the degree of ecosystem isohydricity (σ) is important for predicting plant mortality during drought. However, the current σ estimation method lacks a physiological basis by not explicitly accounting for the photosynthesis process. Recent advances in observing solar-induced chlorophyll fluorescence (SIF), an effective proxy for photosynthetic rate, provides new potential to estimate σ at regional and global scales. Based on the revised mechanistic light response (rMLR) model, we developed a mechanistic, SIF-based model of ecosystem-scale isohydricity forced by satellite SIF observations and meteorological datasets. This model was used to estimate daily global ecosystem isohydricity from 2019 to 2020 at a spatial resolution of 0.25°. During the study period, the mean ecosystem isohydricity of global ecosystems was 0.75, suggesting that, on the whole, global terrestrial ecosystems tend to exhibit more anisohydric behavior. Specifically, herbaceous vegetation mostly displaying highly anisohydric behavior, while woody vegetation tended to be more isohydric. Ecosystem isohydricity exhibits clear seasonality, becoming more isohydric in summer. A random-forest analysis revealed that the three most important factors influencing global ecosystem isohydricity are net photosynthesis rate (Anet), canopy height (h), and vapor pressure deficit (VPD), which collectively accounted for 86.1% of the importance. We also found that ecosystem isohydricity reflects a plant-environment interaction, with the contribution of intrinsic hydraulic traits likely exceeding 50%. The proposed model enables us to incorporate plant physiological controls into the estimation of ecosystem-scale σ, thereby enhancing our ability to track plant water-use strategies in response to rising water stress in a changing climate.