Carbon, nitrogen, and phosphorus are the major plant growth elements, and are mainly absorbed by the plant from the soil. Research on the relationships among these elements is critical for understanding the mechanisms driving the status and balance of these elements in various soils of different land use types. However, to date, there have been relatively few studies on the ecological stoichiometry characteristics of soil carbon, nitrogen, and phosphorus in the sub-humid and semiarid Loess hilly region, which is characterized by severe soil erosion and a fragile ecological environment. Vegetation restoration is one of the most important management approaches for controlling soil and water loss, and for improving soil quality. In the sub-humid and semiarid Loess hilly region, Quercus liaotungensis (oak), which dominates the secondary forests, and Robinia pseudoacacia (black locust) plantations are the major natural and planted vegetation types, respectively. In this study, we investigated the soil organic carbon (SOC), soil total carbon (STC), soil inorganic carbon (SIC), total nitrogen (N), and total phosphorus (P) contents throughout the soil profiles of the two forest types. In addition, the relationships among the aforementioned fractions were analyzed to reveal the ecological stoichiometry of forests in the region. Three stands were selected for each forest type, and a 20 m × 40 m representative plot was surveyed in each stand. Soil samples were collected (using a soil auger with a 6 cm internal diameter) at three points diagonally across the plot at depths of 0-10, 10-20, 20-30, 30-50, 50-100, 100-150, and 150-200 cm. The cored samples from the same depth in each plot were mixed, and all samples were then transferred to the laboratory for further analyses. Prior to the analysis, samples were air-dried, ground to powder, and passed through a 0.25-mm sieve. The SOC content was determined using the potassium dichromate-sulfuric acid oxidation method. The N and P contents were determined using the perchloric acid sulfate cooking method. A FOSS-8400 fully automated Kjeldahl analyzer was used for N quantification. The STC and N contents were also determined using an elemental analyzer. The main results were as follows: (1) SOC and STC contents were linearly correlated (P < 0.0001), and the relationship between SOC and STC could be fitted using the same curve for the two forest types. The relative SOC and SIC contents were stable within a specific soil type from the same area, independent of forest type; (2) in general, the SOC:N ratio was approximately 10, but the STC:N ratio increased with increasing soil depth, and stabilized gradually (saturation curve); and (3) the soil N:P ratio decreased with an increase in soil depth (power law curve). These results provide basic information for the clarification of stoichiometry characteristics in relation to vegetation type and soil depth in this region.