<p>The land carries interactions between natural ecosystems and socio-economic systems, making the regulation of land factors to improve ecosystem service (ES) supply–demand relationships a critical challenge. Taking the Wuhan Metropolitan Area as a case study in 2021, this study employs an optimal parameters-based geographical detector and generalized additive models to explore how land use type, structure, and intensity influence ecosystem service supply–demand relationships at both the pixel and county scales in 2021. The results show that ES supply–demand relationships exhibit spatial heterogeneity. Specifically, the supply–demand relationships of six key ES were evaluated, revealing surplus for regulating services in the eastern mountainous regions and deficit for provisioning and cultural services in urban cores. Furthermore, the explanatory power of land factors for ES supply–demand relationships significantly increases with scale expansion. The response curves of ES supply–demand relationships to land use structure and intensity display nonlinear characteristics with thresholds and scale effects. Additionally, seven ecological management zones were delineated by integrating the surplus-deficit patterns of the six ES at the county scale, each exhibiting distinct spatial characteristics. Corresponding multi-scale land use regulation strategies are proposed, which focus on the targeted adjustment of land use structure and intensity to alleviate specific ES deficits in different zones. This approach offers practical insights for formulating multi-level land management policies.</p>