The spatial compactness of arable land is an important objective of agricultural land consolidation and the basic premise for agriculture modernization. There are already some studies involved in the spatial compactness of arable land, however, most of these studies applied multiple indicators mainly from the perspective of landscape ecology, which could not eliminate multicollinearity among adoptive indicators. In order to gain accurate spatial compactness value and its influence factors, this paper presented a new method with a single index. Firstly, we defined the formulation of spatial compactness from shape, area and location of arable land patches and calculated it with the program code in Matlab; Secondly, we deemed spatial compactness value of farmland patch as dependent variable, and applied econometric method to determine its influence factors; Lastly, Suining County in Jiangsu Province was taken as a case study. Results indicated that the average value of arable land spatial compactness in Suining County was 0.567 7 in 2013, when the radius was set at 2 000 meters, suggesting a high compactness level and good fitness for agricultural modernization. The selected factors of DEM, slope, per capita GDP, per capita farmland, distance to town, density of road network and basic farmland construction affected the arable land spatial compactness obviously, of which slope and basic farmland construction had high marginal function to compactness, while the factor of distance to town had the least marginal function. The proposed method presented in this research could provide technical assistance for basic farmland planning and agricultural land consolidation.