基于RS、GIS以及地理探测器，结合社会经济统计资料，对甘肃农牧交错带2000—2020年土地利用结构、动态度、利用程度和地类转移进行分析，并探究土地利用变化在不同时期的驱动因素。结果表明：（1） 2000—2020年，甘肃农牧交错带以耕地、林地和草地为主要用地类型，三者占比超过95%，土地利用结构差异度均小于0.01、结构总体趋于稳定，全域土地利用动态度总体呈先增加后放缓的趋势；（2） 耕地、林地和草地的转化占研究区总体变化面积的96%以上，各地类间转换活跃，LUCC较集中分布于陇中黄土丘陵区和陇东黄土高原沟壑区；（3） 在长时序土地利用变化情景下，地理探测器具有较好的适用性，甘肃农牧交错带土地利用变化在不同时段驱动因子相异，其变化受社会经济因素和自然地理因素共同作用驱动（前者对土地利用变化解释度更强），各因子间交互作用均表现为双因子增强或非线性增强效应，双因子交互对土地利用变化的解释度强于单因子解释度。

This study analyzed the land use structure, dynamics, intensity, and land use conversion of the agro-pastoral ecotone of Gansu Province from 2000 to 2020 using remote sensing (RS), geographic information system (GIS), and Geodetector in combination with social and economic data. (1) From 2000 to 2020, cultivated land, forest land, and grassland were the main land use types in the agro-pastoral ecotone of Gansu Province, and the total proportion of the three was more than 95%. The difference degree of land use structure was less than 0.01, and the structure was generally stable. The overall land use change initially increased and subsequently slowed down over time. (2) The conversion of cultivated land, forest land, and grassland accounted for more than 96% of the total conversion area in the agro-pastoral ecotone of Gansu Province. Land use conversion was observed among all land use types. Converted cultivated lands were mainly converted to forest land and grassland, whereas other land use types were converted to construction land. Land use type changes were mainly concentrated in the Loess hills in central Gansu and the Loess Plateau gully region in eastern Gansu. (3) In the context of long term land use changes, Geodetector demonstrated good applicability. The driving factors for land use changes in the farming-pastoral ecotone of Gansu varied during different periods. These changes were driven by a combination of socioeconomic factors and natural geographic factors, with the former having stronger explanatory power for land use change. Interactions among these factors exhibited either two-factor enhancement or nonlinear enhancement. Two-factor interactions had stronger explanatory power for land use changes than single-factor explanations.