In the present study, a combination of classification and regression tree (CART) and cluster analysis method was applied in assessing the farmland productivity in Fengqiu County, Henan Province, based on county-level soil spatial and attributive databases. The results indicated that the prediction accuracy of the proposed combination model was considerably improved (to 93.56%) as compared to that by CART approach alone. According to the resulting grading rules, the first, second and third grade farmland accounts for 28.167%, 49.518% and 9.389% of the total area of farmland in this county, respectively; while the fourth and fifth grade farmland accounts for only 5.77% and 7.156% of the farmland, respectively. The higher grading land was mainly distributed in the northwest of Fengqiu County, while the lower grading land was mainly located in the southeast region. There was also an obvious banded decreasing trend of farmland productivity extending from the northwest to the southeast. The results of this paper may help to analyze the spatial distribution of the middle to low-yield fields and limiting factors for improving grain yield, and may also provide references for decision-making on regional farmland management.