Quantifying soil respiration (R_s) and its components [autotrophic respiration (R_α) and heterotrophic respiration (R_h)] in relation to forest management is vital to accurately evaluate forest carbon balance. Thus, R_s, R_α, and R_h were continuously monitored from November 2013 to November 2016 in Pinus massoniana forests subjected to four different management practices in China. We hypothesized that understory removal and thinning decrease R_α and R_h and thus R_s, and these decreases will change with time following UR and thinning. Mean values of R_s, R_α, and R_h in light thinned plots (LT=15% of tree basal area thinned) and heavily thinned plots (HT=70% of tree basal area thinned) were significantly higher than in control (CK) and understory removal plots (UR). The annual R_h/R_s ratio ranged from 58% to 70% across all treatments, and this ratio was significantly higher in HT and LT than in UR and CK. Only HT significantly increased soil temperature. Soil temperature could better explain R_h (R² = 0.69–0.96) than R_α (R² = 0.51–0.86). HT and LT increased Q₁₀ for both R_α and R_h, except for R_h in UR. Soil moisture content (W; %) was significantly higher in HT than in other treatments, but W had limited effects on soil respiration in that rain-rich subtropical China. This result suggests that global warming alone, or in combination with clear-cutting or canopy tree thinning will markedly increase soil heterotrophic respiration and thus the total soil CO₂ emission. To get firewood for local people and to reduce soil CO₂ emissions under global warming, canopy trees are needed to be protected and understory shrubs may be allowed to be used in the subtropical China.