Context Managing land use to promote an ecosystem service (ES) without reducing others is challenging. The spatial scale at which no-loss constraints are imposed is relevant.Objectives We examined the influence of the spatial scale of no-loss constraints on ESs when one ES was optimised. Specifically, we investigated how carbon sequestration could be maximized at different spatial scales in France with constraints of no-loss on other ESs.Methods We used a statistical model linking land use and land cover variables to ESs [carbon sequestration (CS), crop production (CP), livestock production, timber growth] in French small agricultural regions (SARs). We optimised CS at the country scale posing no-loss constraints on other ESs at increasing spatial scales, i.e., SARs (scenario 'SARs'), department ('DEP'), administrative region ('REG'), and France ("FRANCE"). We analysed differences between optimized and initial configurations.Results Optimized CS at the country scale increased with the spatial scale at which no-loss constraints were posed (similar to+ 0.51% for 'DEP' and similar to+ 2.05% for 'FRANCE'). The variability of ES variation among the SARs similarly increased. This suggested that constraints at larger scales lead to ES segregation. Correlations among ES variations changed with the scenarios (Spearman's rho between CS and CP was - 0.43 for 'DEP' and - 0.70 for 'FRANCE'). This indicated that different land use strategies produce different degrees of enhancement/softening of ES trade-offs/synergies. Conclusions A trade-off was highlighted: larger spatial scales promoted better performance of the target ES but also spatial inequality. We argue that addressing smaller scales will lead to land-sharing solutions that avoid the local environmental impacts of land-sparing strategies.