With the Southern Loess Plateau as the object of study, we select the nonbiological factors (physical factors), biological factors and human factors that affect the landscape of arable land to build indicator system. Using GIS, we perform the visualization expression and hierarchical storage of influencing factors to build 1 km×1 km integrated vector and raster database of arable land landscape pattern and its influencing factors. Using spatial regression analysis, we determine the quantitative relationship between arable land landscape and its influencing factors. The results show that the arable land in the Southern Loess Plateau is mainly distributed in the regions with high temperature, great average annual precipitation, high altitude, high soil N content, small slope, GDP per unit area of land, low ≥10℃ accumulated temperature, and short distance away from the rivers and roads. The study provides a scientific basis for clarifying the relationship between arable land landscape and its influencing factors.