[Objective] The characteristics of cultivated land use transformation and the differences in spatial and temporal distribution of carbon emissions were analyzed to explore the mechanisms of influencing factors in the process of cultivated land use transformation carbon emissions, in order to provide scientific basis for the low-carbonization of cultivated land use. [Methods] Using land use data and the carbon emission coefficients, the temporal evolution characteristics of cultivated land use transformation and its carbon emissions from 2000 to 2020 in Hunan Province were shown at the county and 5 km grid scales. The spatial distribution characteristics of carbon emissions were identified and analyzed using the spatial autocorrelation method. The depth of the influence of factors and the degree of interaction between them were assessed using the optimal parameter Geodetector model. [Results] ① Throughout the study period, the area of cultivated land in Hunan Province decreased continuously. With frequent conversions between cultivated land and forest land, water areas, and construction land, carbon emissions showed a fluctuating and increasing dynamic change trend. ② A spatial autocorrelation analysis showed significant high-high and low-low concentration phenomena at both scales, indicating strong concentration distribution characteristics in both spatial and temporal dimensions. ③ The interactions among factors were stronger than those of single factors, with socioeconomic factors and locational conditions having significant explanatory power at both scales. [Conclusion] The effect of cultivated land use transformation on carbon emissions in Hunan Province is significant, with clear differences in carbon emission characteristics and drivers at different spatial scales. Future measures for carbon emission reduction measures and cultivated land use planning should be formulated both scientifically and rationally.