Abstract: The paper studies the spatial differentiation characteristics of agricultural land productivity of the second farmland district in Guangdong province. Through spatial modeling and interpolation, the authors explore the relation between agricultural land productivity and GDP in order to provide scientific support for decision-making of land policy. The result indicates that the agricultural land productivity in Guangdong province is significantly affected by the cultivated land spatial distribution patterns; the theoretical and actual yield of farmland are largely determined by land quality; the correlation between GDP per capita and farm land productivity is different in space. The order of correlation between GDP per capita and agricultural land productivity of the second farmland district in Guangdong province from high to low is in the following: the eastern coastal area and central of Leizhou peninsula, the south central area and the north area, the southwest area and Chaoshan plain area, and Pearl River plains area.