Studies on the effects of artificially restored vegetation on soil nutrient distribution are important to understand the relationship between vegetation restoration and soil environment evolution. However, studies for accurately assessing the ecological benefits of different vegetation restoration models are relatively limited. In this study, we examined the vertical distribution of soil nutrient contents and its influencing factors beneath five typical artificial vegetation types (i.e., Populus simonii, Pinus tabuliformis, Caragana korshinskii, Medicago sativa, and Stipa bungeana) in a transition zone from desert to loess in North China. We found that soil nutrient content had a distinct hierarchy and decreased with increasing soil depth, thereby reflecting the surface aggregation effect of vegetation on soil nutrients. The soil quality index (SQI) and soil nutrients showed the same vertical distribution trend. The average SQI of different vegetation types followed the order: M. sativa (0.2052) > S. bungeana (0.1917) > P. tabuliformis (0.1761) > C. korshinskii (0.1473) > P. simonii (0.0883). Comprehensive analysis showed that the artificial restoration of S. bungeana vegetation was the most suitable restoration type in this catchment. Soil nutrient content was positively correlated with soil water content, plant coverage, clay, silt, and litter nutrient content but negatively correlated with soil bulk density, sand, and litter C/N. Our research is helpful in scientifically evaluating the ecological effect and service value of “Grain-for-Green” program and providing a reference for the implementation of a new round of “Grain-for-Green” in the desert–loess transition zone.