Abstract: Interactions among ecosystem services (ESs) are spatially heterogeneous, shaped by both stable (unidirectional) and unstable (context-dependent) influences of socio-economic development and climate change. These complexities pose significant challenges for spatially adaptive ecosystem management. Ecosystem service bundles (ESBs), as recurring combinations of ESs, offer a valuable framework to capture such interactions. Here, we proposed an integrated analytical framework that combines K-means clustering, the geographical detector (GD) model, and geographically and temporally weighted regression (GTWR) models to evaluate how socio-economic and climatic drivers influence ESBs across space and time. Using Hanshui River Basin (HRB) in central China as a case study, we quantified ES dynamics from 2000 to 2020, identified ESBs and stable and unstable effects of multiple drivers. The results showed that carbon sequestration (CS), water yield (WY), and aesthetic landscape (AL) increased by 32.80%, 9.00%, and 7.34%, respectively, while soil retention (SR) and food supply (FS) declined by 8.38% and 5.36%. Five distinct ESBs were identified, namely ecologically fragile bundle (EFB), agricultural production bundle (APB), water supply bundle (WSB), forest ecological bundle (FEB), and ecological conservation bundle (ECB). Among these, FEB expanded to more than 35.75% of the HRB, and APB exhibited the sharpest decline (–28.64%). Land use intensity (LUI) was the primary driver of the spatial heterogeneity of ESBs, while synergistic and nonlinear interactions among multiple factors increasingly amplified their effects over time. Notably, annual precipitation (AP) emerged as the only stable basin-scale driver, consistently enhancing ES performance, while gross domestic product (GDP) and normalized difference vegetation index (NDVI) had stable yet spatially differentiated effects across bundles. Our findings highlight the significance of distinguishing stable/unstable driver effects on ES dynamics to inform regionally adaptive ecosystem governance. The proposed framework provides valuable insights into ES interactions, identify spatial priorities, and support policy interventions that balance ecological conservation with socio-economic development.