随着全球气候变化对人类社会与经济发展的不利影响日益显著，碳排放成为国际社会普遍关注的问题。农田生态系统作为关键的碳排放源，对其深入研究对制定针对性的减排策略具有重要的指导意义。研究聚焦于2000-2020年中国农田生态系统碳排放的时空格局及其驱动因素。通过分析化肥、农膜、灌溉等农业生产活动及水稻种植过程中的碳排放，揭示了中国农田生态系统碳排放量的变化趋势与空间分布特征。研究采用随机森林算法和累积局部效应模型，对农田碳排放的主要驱动因素进行深入探讨，并通过模拟调控探究了社会经济因素对碳排放的潜在影响。结果显示，中国农田生态系统碳排放量总体呈现先增后减趋势，化肥、农膜和灌溉为主要碳排放源。自然因素、经济因素和社会因素对碳排放量影响显著，且各地区影响过程存在差异。应加强对农业结构和管理方式的优化，制定适宜各地特点的农业减排策略，促进农业生产与环境保护的双赢。

The increasingly significant adverse effects of global climate change on human society and economic development have made carbon emissions a common concern for the international community. The in-depth study of farmland ecosystems is crucial for formulating targeted emission reduction strategies. This study focuses on the spatial-temporal pattern and driving factors of carbon emissions from farmland ecosystems in China during 2000-2020. By analyzing carbon emissions from agricultural production activities such as fertilizer, agricultural film, irrigation, and rice planting, we revealed the changing trend and spatial distribution characteristics of carbon emissions in China's farmland ecosystem. We used the stochastic forest algorithm and cumulative local effect algorithm to deeply explore the main driving factors of farmland carbon emissions, as well as to explore the potential impacts of socio-economic factors on carbon emissions through simulation and regulation. The results showed that carbon emissions from farmland in China increased first and then decreased, with fertilizer, agricultural film, and irrigation being the main sources of carbon emissions. Natural factors, economic development, and social factors have significant impacts on carbon emissions, with differences in influencing mechanisms among different regions. It is important to strengthen the optimization of agricultural structure and management mode while formulating agricultural emission reduction strategies suitable for local characteristics to promote a win-win situation between agricultural production and environmental protection.