Surface ozone (O₃) alters terrestrial carbon–water–energy fluxes. Stomatal flux-based damage schemes nevertheless diverge widely in both impact magnitude and implied carbon–water coupling. Within a unified land surface framework for China, joint comparisons of scheme effects on carbon, water, and energy responses remain limited. Here we couple four stomatal flux-based O₃ damage schemes (Lombardozzi13, Li24, Sitch07, and Ma23) into the Community Land Model version 5.0 (CLM5.0). Using ChinaFLUX sites and regional simulations over China, we compare the schemes at the site, regional, and plant functional type (PFT) scales. The schemes mainly alter the amplitude of simulated fluxes, with little effect on their correlation with observations. Damage intensities differ substantially among schemes. At both site and regional scales, Lombardozzi13 produces the strongest suppression of gross primary productivity (GPP) and transpiration, about 2.4–3.5 times that of the other schemes. Spatial patterns of O₃ impacts are broadly consistent across schemes. Impact fields from Lombardozzi13 and Li24 are more continuous, whereas those from Sitch07 and Ma23 are more concentrated in localised high-value zones. Scheme differences are largest for forest PFTs. Lombardozzi13 overestimates forest damage overall. Li24 suppresses GPP and transpiration in broadleaf forests more strongly than Sitch07 and Ma23, while its responses in needleleaf forests and shrubs resemble those of the latter two. The schemes also alter simulated carbon–water coupling. Li24 yields asynchronous GPP and transpiration responses across vegetation types. By contrast, Sitch07 and Ma23 produce more synchronous responses across most PFTs, except that Ma23 increases grassland sensitivity. These results show that scheme choice controls the magnitude, spatial expression, and carbon–water coupling of ecological O₃ effects. It is therefore a major source of uncertainty in assessing ecological O₃ impacts.