Objective Ecological environmental quality is a fundamental component of regional sustainable development and ecological civilization. Given that most of the existing studies rely on linear or single-scale models, which are difficult to reveal the nonlinear response and spatial heterogeneity characterizing ecological processes, and pay less attention to the long-term dynamic evolution of driving mechanisms. Therefore, this study aims to establish a dual-model coupling analytical framework to systematically analyze the spatiotemporal evolution of ecological quality in mountainous cities and its driving mechanisms, providing scientific support for differentiated ecological governance.

Methods Taking Nanping City in Fujian Province as a case study, this research evaluates the spatiotemporal evolution of ecological quality from 2000 to 2023 based on Google Earth Engine (GEE). The XGBoost–SHAP model is then applied to quantify the nonlinear contributions, threshold effects, and interactions of driving factors. Finally, Multiscale Geographically Weighted Regression (MGWR) is used to reveal the spatial heterogeneity and differences in the action scales of these factors.

Results (1) From 2000 to 2023, the ecological quality of Nanping City showed a fluctuating upward trend, with the mean RSEI increasing from 0.649 to 0.671. Spatially, a pattern of lower values in the central area and higher values in the surrounding areas was observed. (2) The ecological quality of Nanping City is mainly improved during 2000—2023, and its area share is 41.62%. In contrast, areas with ecological degradation were mainly concentrated in Jianyang District, Shaowu City, and Yanping District, covering 33.56% of the study area. (3) The XGBoost-SHAP results indicate that the driving mechanism of ecological quality in Nanping City shifted from being dominated by natural factors to a combined pattern involving both human activities and natural background conditions between 2000 and 2023. In particular, human-related factors such as the proportion of cropland and built-up land showed a marked increase in importance and became key negative drivers of ecological quality. (4) The MGWR results further reveal that the effective scales of different driving factors are not uniform and that their spatial influence scales change over time. Among these factors, elevation, slope, and population density exhibit the most pronounced spatial heterogeneity.

Conclusion This study identifies the key driving factors of ecological quality changes in Nanping City and clarifies their threshold values and spatial scales by the XGBoost-MGWR dual-model coupling framework. The findings provide theoretical and data-based support for ecological spatial optimization and targeted governance in Nanping City and other mountainous cities.