The mechanisms behind changes in the trade-offs between ecosystem services (ESs) over time and space have been the subject of extensive research. However, studies on the drivers influencing ES trade-off changes across various climate zones are limited. This study examines the drivers that determine the ES trade-off changes in the Sichuan-Yunnan-Loess Plateau ecological shelter (SYLPES), China, across several climate zones. Three ESs, namely water yield (WY), soil conservation (SC), and net primary productivity (NPP) were evaluated using the Integrated Valuation of Ecosystem Services and Trade-offs model, the Revised Universal Soil Loss Equation model, and the Carnegie-Ames-Stanford Approach model. The trade-offs between ES pairs were quantified by the root mean square error. In addition, we used the geographic detector model to investigate the drivers of ES trade-off changes. The findings indicate spatial heterogeneity in all ESs. Low values were observed in the north-eastern region of SYLPES and high values in the south-western area. Except for NPP, which decreased slightly in the north subtropical zone (NSZ), all ESs exhibited a general increase across climate zones. Significant spatiotemporal changes were observed for the trade-offs, with WY boolean AND SC increasing and WY boolean AND NPP and SC boolean AND NPP decreasing from the north to south. The highest WY boolean AND SC trade-off occurred in the highland climate zone (HCZ), while the WY boolean AND NPP trade-off peaked in the southern temperate zone (STZ), and the SC boolean AND NPP trade-off was greatest in the central subtropical zone (CSZ). The factors driving changes in ES trade-offs varied across climate zones, with natural drivers, particularly precipitation, playing a significant role. Socioeconomic drivers had a greater impact on trade-off changes in the STZ and medium temperate zone (MTZ) than in the NSZ, CSZ, and HCZ. In the last five years, socioeconomic drivers had a stronger impact on trade-off changes in the MTZ, STZ, and HCZ (especially the MTZ), but a weaker impact on trade-off changes in the NSZ and CSZ (especially the CSZ) than in the previous decade. This study provides a scientific basis for managing ESs that span climate zones on a large scale.