By using Li-6400-09 system, an in situ measurement on the soil respiration, soil temperature, and soil moisture content in three main forest types (evergreen broadleaved forest, Cunninghamia lanceolata plantation, and Phyllostachys pubescens forest) in subtropical area of China was conducted, with the relationships between soil respiration and soil temperature and moisture content simulated by several models. The C flux of soil respiration in P. pubescens forest, evergreen broadleaved forest, and C. lanceolata plantation was 12.84, 11.70, and 7.12 t C x hm(-2) x a(-1), respectively, and the soil respiration in these three forest types had similar diurnal and seasonal variations, with the maximum value at 11:00-12:00 and the minimum value at 1:00-3:00, and the highest value in August and September while the lowest value in December and January. Van't Hoff equation and Lloyd & Taylor function had less difference in describing the relationships between soil respiration and soil temperature, while the soil respiration rate predicted by Lloyd & Taylor function was smaller than the observed value. Quadratic model and power function model could well simulate the relationship between soil respiration and soil moisture content. Soil moisture content positively or negatively affected soil respiration, but the effects only reached significant level in C. lanceolata plantation. Comparing with single-factor equation, two-factor equation (soil temperature and moisture content) could better describe the responses of soil respiration to changed soil temperature and moisture content. Multivariate analysis of covariance showed that after eliminating the effects of soil temperature and moisture content, forest type had significant effects on soil respiration (R2 = 0.541). Other factors, such as air temperature, air relative humidity, and photosynthetic radiation also affected soil respiration, and the effects of air temperature reached significant level.