Abstract: Abstract: Soil erosion is one of the most widespread ecological problems in the world. It accelerates the process of land desertification, causes soil degradation and soil productivity reduction; affects the agricultural development and food security; and seriously restricts the sustainable development of society, economy, and ecology in the world. In the black soil region of Northeast China, where is the most important grain production area and one of six soil erosion regions in China, serious soil erosion causes a decrease in the black soil thickness, leading to a steep decline in soil productivity. In this study, the influence of soil erosion thickness on the indexes of soil physicochemical properties, soybean biological traits and water use efficiency in the black soil region were studied in a pot experiment in 2012 and 2013. The tested soil was obtained from 0, 5, 10, 15, 20, 25, and 30 cm below the surface layer to simulate the plough layer soil with soil erosion thickness of 0, 5, 10, 15, 20, 25, and 30 cm. A quantitative model of soybean yield in the black soil region on the soil erosion thickness was established to explore the response of soybean yield to soil erosion thickness. Combining the physical and chemical properties of the soil in the experimental zone, the TOPSIS model is improved, by which the soil productivity under different soil erosion thickness was evaluated. The results showed that soil erosion thickness had a significant influence on many important indicators of soil quality. Total N, alkaline hydrolysis N, total P, available P, organic matter content, and field moisture capacity decreased, whereas soil bulk density increased. With the increase in erosion thickness; total K and available K contents had no remarkable changes. Soil erosion thickness also had a significant influence on soybean growth, which was mainly reflected by the reduction in plant height, decreased in the number of pods and grains per plant, resulting in decrease of soybean yield. With the increase of the soil erosion thickness, soybean yield reduction rate increased in a S-shaped curve, soybean yield and water consumption decreased in a Z-shaped curve, while water use efficiency decreased in the form of exponential function. The TOPSIS and improved TOPSIS models were used to calculate the soil productivity indexes under different soil erosion thicknesses, and a comparative analysis was carried out between the results of the two models and the commonly used fuzzy matter-element model. The three models had a high consistency in the soil productivity evaluation, and among which the improved TOPSIS model was the most adequate one. The soil productivity indexes calculated by the fuzzy matter-element and TOPSIS models showed a good linear relationship with soil erosion thickness, while the indexes calculated by the improved TOPSIS model showed a Z-shaped curve, which was consistent with the change of soybean yield. To further analyze the evaluation results of the three models, regression analysis was conducted on the soil productivity indexes and the soybean yield. The result showed that, soybean yield had a linear relationship with the soil productivity indexes calculated by the fuzzy matter-element and TOPSIS models but an exponential relationship with that calculated by the improved TOPSIS model. The soil productivity indexes calculated by the improved TOPSIS model developed from this study can properly reflect the soil productivity levels under different soil erosion thicknesses and can be used for the evaluation of soil productivity. The results will provide a more reasonable and reliable evaluation method and valuable information for further study on soil erosion prevention and control in the black soil region.