Abstract: Abstract: There is a dilemma between raising the intensive use level of agricultural land and controlling carbon emissions from agricultural land use. Carbon emissions from agricultural land use can be considered as a cost, and the suitable intensive use level of agricultural land should acquire a good efficiency of carbon emissions from agricultural land use. First, an assessment model for evaluating agricultural land use intensity was established in this article, and the input-output indices of agricultural land intensive use were selected. The input-output indices of agricultural land intensive use were composed of labor input per unit of agricultural land, capital input per unit of agricultural land, technique input per unit of agricultural land, value-added of agriculture per unit of agricultural land, grain products per unit of agricultural land, income from household business per unit of agricultural land, and ammonia-nitrogen emission from agriculture per unit of agricultural land. Secondly, the intensity values of agricultural land intensive use in 31 provinces were computed. Thirdly, based on a slacks-based measure model, the input-output indices were established. The input-output indices applied in the slacks-based measure model were composed of labor input per unit of agricultural land, capital input per unit of agricultural land, technique input per unit of agricultural land, value-added of agriculture per ton of carbon emission, grain products per ton of carbon emission, income from household business per ton of carbon emission, and ammonia-nitrogen emission from agriculture per ton of carbon emission. Fourthly, this article calculated the total efficiency values, technical efficiency values, and scale efficiency values of carbon emission from agricultural land intensive use in 31 provinces and designed an input-output optimization of agricultural land intensive use. Lastly, low-carbon optimization strategies for agricultural land use in provinces whose total efficiencies of carbon emission were inefficient were put forward. The results showed that: 1) Consistency between agricultural land use intensity and its efficiency of carbon emissions does not exist widely in China, and the provinces whose intensity values of agricultural land use are large are mainly located in eastern China and the traditional farming zone. 2) The provinces whose total efficiencies of carbon emission from agricultural land intensive use are DEA efficient were mainly distributed in western China, and the total efficiencies of provinces in the Yangtze Delta, the Zhujiang Delta and Jing-Jin-Ji Megalopolis which are in an economically developed area are widely low. 3) The main reason for the low total efficiency was the imperfect technical efficiency of carbon emissions from agricultural land intensive use, and the imperfect scale efficiency had little effect on the low total efficiency. 4) In order to improve the total efficiency of carbon emissions from agricultural land intensive use, some measures should be adopted according to the local features of agricultural land use and the economy. These measures mainly consist of transferring agricultural labor, controlling carbon sources input, developing low-carbon and high-output agriculture, etc. Effective measures can be applied to control the carbon emissions from agricultural land use, and also can be added to the good outputs which contain added value of agriculture, grain output, and income from increased household business. of course, the emissions of ammonia nitrogen discharged from agriculture can also be reduced by improving agricultural land use.