Abstract: A two-factor randomized block design was adopted to set up a field trial with a total of 9 treatments (3 replicates for each treatment) of two factors (nitrogen and phosphorus, were set up in the field for. After four-year fertilization, the stratified samples (0-10 cm, 10-20 cm, 20-40 cm) were taken to determine the soil nitrogen and phosphorus contents and the activities of related key soil enzymes, as well as the leaf biomass per plant was simultaneously measured. The LASSO regression model and the partial least squares path model (PLSPM) were used to analyze the effects of vertical distribution of soil nitrogen and phosphorus nutrients on leaf biomass. Results showed that under the combined application of nitrogen and phosphorus, the activities of nitrogen and phosphorus nutrients and related enzymes in the soil of the C. paliurus plantations all showed significant surface aggregation (0-20 cm) distribution characteristics in the soil, while the combined application of nitrogen and phosphorus had a significant interaction effect on the leaf biomass. Among them, the treatment with medium nitrogen and high phosphorus (N₁P₂) and high nitrogen and medium phosphorus (N₂P₁) demonstrated the best effect, and the leaf biomass per plant was more than four times that of the control (N₀P₀). The four key soil factors affecting leaf biomass were the alkali-hydrolyzable nitrogen content in the soil at 0-20 cm, the acid phosphatase activity in the soil at 0-10 cm, the available phosphorus content in the soil at 0-10 cm, and the total nitrogen content in the soil at 0-20 cm. In addition, the established model based on the 0-20 cm soil layer had the best fit goodness and the best explanatory ability for leaf biomass production (Gof = 0.64, R² = 0.67).