Soil nutrient stoichiometry is fundamental to ecosystem functioning. However, it remains unclear how long-term vegetation restoration affects its dynamics. This study utilized long-term monitoring to compare soil carbon (C), nitrogen (N), and phosphorus (P) stoichiometric characteristics with cropland as a control, across four stand types: cropland (CK), natural secondary forest (NSF), Robinia pseudoacacia and Pinus tabuliformis mixed forest (RPF), R. pseudoacacia plantation (RF), and P. tabuliformis plantation (PTF). Specific focus was given to soil organic carbon (SOC), total nitrogen (TN), and total phosphorus (TP), as well as their C:N, C:P, and N:P ratios. During 30 years of vegetation restoration, SOC and TN increased by 11.65% and 7.46%, while TP decreased significantly by 11.74%. Furthermore, SOC and TN increased significantly with natural restoration forest (NSF) and stabilized by the 25th year. Under artificial restoration (PF, RF, RPF), SOC and TN initially decreased but recovered to their initial values after approximately 13-15 years, and stabilized by the 25th year. Finally, the C:N ratio did not change significantly, while the C:P and N:P ratios varied temporally with early fluctuations (0-10 year period), significant increases in the mid-term (10-20 year period), and stabilization in the late stage (20-30 year period). Climate indirectly influenced SOC, TN, and TP through vegetation factors. SOC and TN were directly affected by both soil and vegetation factors, while TP was primarily influenced by vegetation factors. Over the 30-year period, as vegetation biomass gradually accumulated, the nutrient dynamics shifted from being soil-mediated to vegetation-mediated. Long-term vegetation restoration primarily enhances soil nutrient dynamics through biological pathways, with trees and herbaceous plants playing key roles in nutrient dynamics and stoichiometric optimization. NSF is more effective than artificial restoration in enhancing soil nutrient accumulation and optimizing stoichiometric ratios. These results offer valuable guidance for developing effective restoration strategies in the Loess Plateau.