It is of practical significance for the accurate evaluation and prediction of the carbon budget of the natural reserve system in the region with national park as main body to explore the dynamic changes of soil respiration (R_s) and the influencing factors under different forest land use patterns in subtropical mountain-type national park. In Wuyishan National Park, the soil respiration, near surface temperature, soil temperature, soil moisture, soil nutrients, soil microbial carbon (MBC) and soil microbial nitrogen (MBN) of tea garden, Castanea henryi (skam) Rehd. et Wils. forest, Pinus massoniana Lamb. forest and bare land were measured by Li-8100 open circuit soil carbon flux measurement system. The results showed that:(1) Same as the near surface temperature, soil temperature and soil moisture, the R_s of different forest lands were ranked seasonally as summer > spring > autumn > winter, and the seasonal mean of the R_s was in the order of tea garden (3.10 μmol m^-2 s^-1) > P. massoniana (2.96 μmol m^-2 s^-1) > C. henryi (2.32 μmol m^-2 s^-1) > bare land (1.43 μmol m^-2 s^-1). There were extremely (P<0.01) significant differences between C. henryi forest and bare land, and between C. henryi forest and P. massoniana forest. Except for P. massoniana forest, the single-factor quadratic polynomial model of hydrothermal factors (near surface temperature, soil temperature and soil moisture) had the highest fitting degree to R_s. In the composite model jointly established by hydrothermal factors, the power-exponential model established of soil temperature and moisture had a high fitting degree to R_s of tea garden. Soil temperature and moisture can explain 80% of the change in R_s. The R_s of P. massoniana forest was suitable for the logarithmic function model established by soil temperature and moisture, and the three-factor linear model (Enter-regression method) had the best fit for R_s in C. henryi forest and bare land, with R² of 0.565 and 0.281, respectively. (2) The contents of the carbon, nitrogen, and phosphorus of tea garden and C. henryi forest were higher than those of P. massoniana forest and bare land. The order of MBN content was tea garden > P. massoniana > C. henryi > bare land. The difference was the biggest in summer, and the range was 59.47 mg/kg. There was an extremely (P<0.01) significant positive correlation between R_s and total phosphorus (TP), available phosphorus (AP), total potassium (TK) and available potassium (AK) content in tea garden. There was an extremely significant (P<0.01) positive correlation between R_s and TP, TK and AK content in P. massoniana forest. There was and a significant (P<0.05) positive correlation between R_s and TK, AK and MBN content in C. henryi forest. The R_s of bare land was significantly (P<0.05) affected by MBN content. The R² of multiple stepwise regression equations in four kinds of forest land are all close to 1. In summary, the soil respiration of tea garden and P. massoniana forest were higher, and the soil respiration of all forest land showed the seasonal dynamics of summer > spring > autumn > winter. The correlation between temperature and soil respiration, and the correlation between moisture and soil respiration were significant. Temperature and moisture were dominant factors in the seasonal change of soil respiration in subtropical mountain area with abundant hydrothermal conditions. Among them, the response of soil respiration to hydrothermal factors in tea garden of Wuyi Mountain was the most sensitive among four forest lands. In addition to temperature and moisture, the soil respiration of various forest lands was greatly affected by P and K elements, among which tea garden was mainly affected by P element, and that of P. massoniana forest was more affected by K element.