Abstract: Abstract: The investigation on carbon dioxide emission efficiency is essential in achieving the environmental and ecological benefits of cultivated land intensive use. However, the specific investigation of the relationship between cultivated land intensive use and carbon dioxide emission efficiency is still limited, especially in the provincial scale. Therefore, the research explored the variation law of carbon dioxide emission efficiency under cultivated land intensive use of different levels in Hebei Province, owing to its essential geographical position in the food security and economic development of the Beijing-Tianjin-Hebei Metropolitan Region. The main methods included the construction of the model of cultivated land intensive use and the model of data envelopment analysis (DEA). The indices were chosen from input (labor, capital and technology per agricultural acreage) and output. As carbon emission was considered as the cost, the agricultural added value, grain production and family business income per unit carbon dioxide emission were determined as output indices. GIS (geographic information system) spatial analysis was adopted to investigate the level of cultivated land intensive use and the features of the temporal and spatial variation of carbon dioxide emission efficiency in 138 counties (cities) in Hebei Province. Exploratory spatial data analysis (ESDA) and Moran scatter diagram were adopted to measure the aggregation degree of cultivated land intensive use and carbon dioxide emission efficiency. On the basis above, the research investigated the features of spatial clustering difference of cultivated land intensive use and total carbon dioxide emission efficiency, and analyzed the variation law of carbon dioxide emission efficiency under cultivated land intensive use of different levels. The results were as follows: 1) The level of cultivated land intensive use in Hebei Province presented the expanding trend from plain to mountain and from hill to Bashang Plateau. 2) The effective areas of total carbon dioxide emission efficiency were mainly centered on the Bashang Plateau and mountain and hill areas, while the total carbon dioxide emission efficiency was relatively low in developed areas such as Shijiazhuang, Xingtai and Tangshan, because of the insufficient utilization of input and the consequent failure of restricting the scale of carbon dioxide emission during the increase of grain production and agricultural added value. 3) Inverse association was presented between the level of cultivated land intensive use and the total carbon dioxide emission efficiency spatially, and the local aggregation effect was constantly reinforced over time. Most areas showed HL (high cultivated land intensity use - low carbon dioxide emission efficiency) clustering and LH (low cultivated land intensity use - high carbon dioxide emission efficiency) clustering, indicating the necessity of investigating the relationship between cultivated land intensive use and carbon emission optimization. The risk of over intensive use existed in the areas of HL clustering, which demanded the protection of ecological benefit. The ecological vulnerability was the bottleneck of the areas of LH clustering, and therefore, its priority should be guaranteed during the optimization of resource utilization. Owing to the limitation of reserved cultivated land resources, the carbon emission should be emphasized during the land resource management in Shijiazhuang City, Tangshan City, Qinhuangdao City, Langfang City and Chengde City. Relatively, Baoding City and Handan City owned sufficient reserved cultivated land resources, thus demanding comprehensive land improvement.