Mineral soil respiration is an important route for the loss soil carbon (C) in forest ecosystems, and it is a key factor for estimating the forest ecosystem carbon balance. Making sense of mineral soil respiration variation on temporal and spatial scales is a critical component of understanding forest ecosystem C cycle responses to climate change. Thus, high-frequency measurements have become one of the primary tools used to measure mineral soil respiration on both temporal and spatial scales. The objective of this study was to compare a manual system (Li-8100A) with an automated system for measuring mineral soil respiration in a subtropical forest. We carried out the study in a secondary Castanopsis carlesii forest with different forest management (Control, CK; Residual retention after clear cutting, RR; Residual burning after clear cutting, RB) in Sanming, Fujian. The results obtained were as follows. (1) The mineral soil respiration flux, soil temperature, and soil moisture content under different forest managements all showed obvious seasonal dynamics, and mineral soil respiration rates were in the order of CK (2.18 μmol m^-2 s^-1) > RB (1.93 μmol m^-2 s^-1) > RR (1.89 μmol m^-2 s^-1). (2) The annual flux determined using high-frequency measurement was significantly higher than that determined using the manual system, whereas there was no difference between the soil respiration rate calculated between 09:00 and 11:00 (high-frequency data) for different forest managements. (3) Soil hydrothermal condition is one of the important factors affecting mineral soil respiration under different forest management. A two-factor model show that included soil temperature and moisture explained the annual variation of mineral soil respiration rate as 96.8%, 62.8% and 95.4% in CK, RR, and RB treatments, respectively. The double factors model demonstrated that combined with soil temperature and moisture was explained the annual variation of mineral soil respiration rate 96.8%, 62.8% and 95.4% in CK, RR and RB, respectively. The results obtained using this model were better than those obtained using a single-factor model. Therefore, in order to accurately assess and predict the effects of different forest management on the C flux exchange between the soil and atmosphere under future climate change, high-frequency measurements should be considered.