Exploring the characteristics, influencing factors and trend forecast of carbon emission from farmland inputs, and putting forward the policy proposal to promote the carbon reduction of farmland inputs has great significance to the transformation of low carbon and green agriculture. In this paper, the growth mechanism and the trend forecast of carbon emission from farmland inputs in Hunan Province were studied by Regression Analysis, STIRPAT and GM (1,1) model. The results show that the annual growth rate of carbon emission from farmland inputs in Hunan Province increased 3.25% from 2000 to 2014. The total carbon output of chemical fertilizers in various carbon sources was the highest, followed by pesticide; The average annual growth rate of total carbon emissions from tillage, pesticide, agricultural film, diesel, irrigation and chemical fertilizer was 0.68%, 3.22%, 7.47%, 6.42%, 2.57% and 3.25% respectively. The agricultural population, per capita agricultural GDP, mechanization level, agricultural production efficiency, and agricultural industry structure significantly affected the carbon emissions from farmland inputs, and each change of 1% will bring about 0.20%, 0.95%, 0.12%, 0.98% and 0.93% change in carbon emissions from farmland inputs. The GM (1,1) model was used to forecast the carbon emissions from farmland inputs in Hunan Province from 2016 to 2020, and the carbon emission was rising continuously, which is expected to 430.43 million tons in 2020. Finally, according to the conclusions of the study, it puts forward some policy suggestions to promote the carbon reduction of farmland inputs.

Key words：farmland inputs; carbon emissions; STIRPAT Model; GM (1,1); Hunan Province