Assessing the emission-peaking process of agricultural land use provides valuable insights for mitigating global warming. This study calculated agricultural land-use carbon emissions (ALUCEs) in China from 2000 to 2020 and explored the peaking process based on quantitative criteria. Further, we applied the Tapio decoupling index and environmental Kuznets curve (EKC) model to discuss the robustness of the peaking process. The main conclusions are as follows: (1) From 2000 to 2020, China’s average ALUCEs were 368.1 Mt C-eq (1349.7 CO₂-eq), peaking at 396.9 Mt C-eq (1455.3 Mt CO₂-eq) in 2015 before plateauing. Emissions from agricultural materials and soil management had entered the declining period, while those from rice cultivation were in the peaking period, those from straw burning were still rising, and those from livestock breeding remained at the plateauing phase. (2) The provinces of Beijing, Tianjin, and nine others saw a decline in ALUCEs, while Hainan, Guizhou, and another nine provinces observed plateauing, and Ningxia, Qinghai, and six other provinces experienced peaking. (3) Decoupling analysis confirmed that emission-peaking states remained stable even with agricultural growth. Instead of an inverted U-shaped relationship, we found an N-shaped relationship between ALUCEs and agricultural GDP. The spatial EKC model indicated that the peaking process had spillover effects between provinces. It is recommended that China accelerate ALUCE mitigation based on the source and phase of emissions, considering the peaking process and magnitude.