Evaluation of Cultivated Land System Health and Its Driving Factors in the Dabie Mountains Using a TOPSIS–GRA Model

Cultivated land health is fundamental to regional food security and sustainable agricultural development, particularly in hilly and mountainous areas where land fragmentation and natural constraints are pronounced. Taking Macheng City in the Dabie Mountains as the study area, this study constructed a 23-indicator cultivated land health evaluation system based on the Driving Force–Pressure–State–Impact–Response (DPSIR) framework using multi-source spatial and socioeconomic data for 2014 and 2024. Indicator weights were determined by the Criteria Importance Through Intercriteria Correlation (CRITIC) method, and cultivated land health was evaluated using a TOPSIS–GRA coupling model. Hot spot analysis and the geographical detector were used to identify spatiotemporal patterns and driving factors. The combined proportion of healthy and very healthy cultivated land increased from 54.08% to 58.92%, although the five DPSIR subsystems showed divergent trends. High-value clusters were concentrated in western plains and river valleys, while persistent cold spots occurred in eastern mountainous areas. Agricultural mechanization was the dominant driver in 2024 (q = 0.4946). The driving mechanism shifted from labor and material inputs toward the joint influence of mechanization and natural background conditions. These findings provide empirical evidence that improvement in an aggregate cultivated land health index does not necessarily indicate balanced improvement across the underlying subsystems, offering a case-based reference for mountain–hill agricultural regions undergoing similar transitions.