The Loess Plateau is the area with the most serious soil erosion, and soil restoration is urgently needed. As the basic unit of soil erosion, slope is of great significance to the study of soil erosion factors, and slope vegetation is an effective means to improve soil conditions. To remediate the slope soil on the Loess Plateau, a mixed vegetation restoration model was used to remediate the soil in this study. In the experiment, the factors related to soil remediation of different vegetation patterns on the slope scale were quantitatively analyzed, and the configuration optimization of slope mixed vegetation coverage, relative distance and patch pattern was carried out. When the rainfall intensity is 55 mm/h, the best coverage when the slope is below 15 degrees is 69.3%; the relative distance is better when it is 0.2, and the optimal relative position for runoff and sediment reduction is 0.0-0.22. When the connectivity of mixed vegetation patches is better, the density of patches is greater, and the shape of patches is more complex, the runoff and sediment yield rates are lower, and the sediment accumulation is significantly reduced. Based on the results, the key indexes that affect slope soil remediation were screened out, and the optimized pattern of slope vegetation was obtained. Through data analysis, the erosion mechanism of the Loess Plateau is clarified, which provides data support for soil restoration in this area. At the same time, the optimal pattern of vegetation is given in the experiment, which provides a specific model for soil remediation in the Loess Plateau, and is conducive to soil remediation in the Loess Plateau.