In order to reveal the temporal and spatial differentiation characteristics of carbon emissions from cultivated land use transition, taking Harbin section of Songhua River Basin as an example,the carbon emission intensity of cultivated land use transition from 1990 to 2020 was measured based on the grid element method and carbon emission coefficient. In addition, the means of gravity center analysis, exploratory spatial data analysis, cold and hot spot analysis tools and geodetector were implemented to reveal the spatial heterogeneity and formation mechanism of carbon emissions from cultivated land use transition. The results showed that the cultivated land area demonstrated a downward trend in 1990—2020, and the conversion between cultivated land and construction land, forest land was the most dramatic, and cultivated land mainly converted into construction land.The carbon emission of cultivated land use transformation showed an upward trend in 1990—2020, from 3704.12t in 1990—2000 to 35656.29t in 2010—2020 with an increase of nearly 8.63 times. The carbon emission of cultivated land use transformation eventually appeared as a carbon source. From 1990 to 2020, the carbon emission center of arable land use transformation remained basically stable, moving 15.17km to the east, of which the center of gravity moved the longest from 1990 to 2010. Overall the carbon emissions in the northeast region continue to deteriorate, while the carbon emissions in the southwest region were improved significantly. In 1990—2020, the carbon emissions from cultivated land use transformation were non randomly distributed, but had strong spatial agglomeration. The hot spots mainly spread around Nangang District to the surrounding boundary, and the cold spots were scattered in the southeast. Urban center distance was the main factor of carbon emission for cultivated land use transformation, and the interaction between various factors was mainly enhanced by two factors. Among them, the interaction between land use degree, and annual average precipitation, urban center distance had strong explanatory power. The spatial differentiation characteristics of carbon emission for cultivated land use transformation were the result of multi-factor interaction.