Exploring the differential responses of rhizosphere soil phosphorus contents associated with nitrogen-fixing and non-nitrogen-fixing plants to different soil nitrogen levels in subtropical karst forests can provide valuable insights into the effects of nitrogen-fixing plants on soil nutrient cycling. Such knowledge will serve as a scientific reference for the extensive planting of nitrogen-fixing plants in vegetation restoration efforts in karst regions. Taking karst forests with varying soil nitrogen levels in Jianshui County, Yunnan Province as test objects, we collected soil samples from the rhizosphere of three types of dominant nitrogen-fixing and non-nitrogen-fixing plants with the same age and analyzed the total phosphorus (TP), organic phosphorus (OP), inorganic phosphorus (IP), available phosphorus (AP), and other soil physicochemical properties. Soil microbial biomass and enzyme activities were measured to assess the influence of nitrogen-fixing plants on rhizosphere soil phosphorus contents under different soil nitrogen levels, as well as the main driving factors. Results showed that the contents of TP, OP and AP in the rhizosphere soil of nitrogen-fixing plants significantly increased by 16.0%, 66.5% and 139.5% under a low soil nitrogen level with the available nitrogen of 15.62 mg·kg, and significantly increased by 13.5%, 25.7% and 15.7% under higher soil nitrogen level with the available nitrogen of 37.15 mg·kg, respectively. There was no significant difference in IP content between nitrogen-fixing and non-nitrogen-fixing plants under the two soil nitrogen levels. Compared with low soil nitrogen level, the contents of TP and IP in the rhizosphere soil of nitrogen-fixing plants under high soil nitrogen level significantly decreased by 21.3% and 31.7%, and those of non-nitrogen-fixing plants significantly decreased by 19.6% and 39.1%. The AP content in the rhizosphere soil of nitrogen-fixing and non-nitrogen-fixing plants significantly increased by 32.8% and 174.8%, respectively, with no notable change in OP content. Under low nitrogen conditions, nitrogen-fixing plants significantly increased microbial biomass carbon (MBC), microbial biomass nitrogen (MBN), microbial biomass phosphorus (MBP), and alkaline phosphatase (ALP) activity in the rhizosphere soil. Under high nitrogen condition, nitrogen-fixing plants significantly increased MBP and ALP activity, but had no significant effect on MBC and MBN. As soil nitrogen level increased, soil MBC, MBN, MBP, and nitrogen cycle-related enzyme activities in the rhizosphere soil of nitrogen-fixing plants decreased significantly, while ALP activity increased. In contrast, in the rhizosphere soil of non-nitrogen-fixing plants, MBN and ALP activity significantly increased, while nitrogen cycle-related enzyme activities significantly decreased. Mantel analysis indicated that under low nitrogen level, rhizosphere soil phosphorus contents were primarily regulated by a combination of soil physicochemical properties, microbial biomass, and enzyme activity, while they were mainly regulated by soil physicochemical properties under high nitrogen level. In conclusion, compared to non-nitrogen-fixing plants, nitrogen-fixing plants in subtropical karst forests can significantly increased soil TP, OP, and AP contents and this effect is largely regulated by soil nitrogen level. Therefore, introducing nitrogen-fixing plants into low-nitrogen subtropical karst areas at the beginning of vegetation restoration may alleviate phosphorus limitation, improve soil nutrient status, and facilitate vegetation restoration in these regions.