Objective This study aims to reveal the spatiotemporal differentiation patterns of cultivated land landscape from the provincial scale to economic zones in Sichuan Province， and to quantify the contributions and interactive effects of natural and socio-economic factors on landscape pattern changes. Methods Based on multi-source remote sensing and geographic information system （GIS） technology， a spatiotemporal evolution model of landscape ecology was constructed on the ArcGIS and Fragstats platforms. Combined with the geodetector and entropy weight-TOPSIS method， the spatial heterogeneity characteristics and driving mechanisms of cultivated land landscape patterns in Sichuan Province and its five major economic zones from 2000 to 2022 were systematically analyzed. Results 1） From 2000 to 2022， the cultivated land in Sichuan Province generally showed a trend of scale reduction， fragmentation intensification， and connectivity decline. The area of dryland decreased by 13 961.0 km²， with a higher reduction rate than that of paddy fields. The number of patches increased sharply， and landscape dominance continued to weaken. 2） The five major economic zones exhibited significant differences. The Chengdu Plain Economic Zone showed quantity reduction and shape fragmentation. The Western Sichuan Economic Zone presented paddy field expansion and dryland shrinkage. Both paddy fields and dryland in the Northeast Sichuan Economic Zone decreased by more than 19%. The area of dryland in the Southern Sichuan Economic Zone decreased by 11.58%， with local retention of core cultivated areas. The area of dryland in the Panxi Economic Zone decreased by 27.49%， with the highest degree of fragmentation. 3） In terms of driving mechanisms， the interactions between natural and socio-economic factors were dominated by two-factor enhancement and nonlinear enhancement effects. Among them， population density had the highest interactive explanatory power （0.97）. The influence of human-driven factors generally increased across regions. Economically developed areas were mainly driven by urbanization and industrialization， ecologically sensitive areas were dominated by natural constraints and policy regulation， and traditional agricultural areas were trapped in a cycle of agricultural decline and population loss. Conclusion The findings of this study can provide a theoretical basis for the construction of an ecological security pattern for cultivated land and differentiated management strategies in Sichuan Province， thereby supporting the sustainable use of cultivated land landscapes and the process of territorial spatial governance.