With the continuous development of society, the agricultural economy has shown a good development trend, but at the same time it has produced more and more carbon emissions, and the greenhouse effect has increased, posing a threat to the ecological environment. Agricultural carbon emissions are important factors. Therefore, studying the temporal and spatial characteristics of agricultural carbon emissions is an urgent task. This paper focused on agricultural carbon emissions from land use and tested and predicted future carbon emissions based on three types of carbon source factors using the auto regressive integrating moving average (ARIMA) model. The spatial and temporal characteristics of agricultural carbon emissions were analyzed from the dimension of time and space. According to the results, the total agricultural carbon emission and intensity of the eastern region ranked the first, followed by the middle region; the total agricultural carbon emission of different regions increased, but slower; the agricultural carbon emission intensity of different regions decreased, and overall decreased faster. This study effectively verifies the deep influence of the development of ecological agriculture on the spatial and temporal characteristics of carbon emission of land use.