Soil respiration, the second largest CO2 flux in terrestrial ecosystems, involves the microbial respiration of litter from aboveground sources, belowground litter and root respiration, from rhizodeposition. In the subtropical forests of the Ailao Mountain, a chamber with automated CO2 efflux was set up with two treatments: a control treatment with litterfall to measure the total soil respiration (RT) and a litter removal treatment to measure aboveground litterfall respiration (RL). This study aimed to examine the responses of RL to unexpected heavy snowfall events and soil temperature (ST), soil moisture (SM), rainfall (RF), total litterfall (TL), litter water content (LWC), nitrate nitrogen (NO3−-N), and ammonium nitrogen (NH4+-N). The period of the study was divided into two: before the heavy snowfall event (BS) and during and after the heavy snowfall event (AS). The rate of RL was slightly decreased in AS (1.18 ± 0.03 μmol CO2 m−2 s−1) compared with that in BS (1.19 ± 0.02 μmol CO2 m−2 s−1). The relationships between RL and ST, SM, RF, and LWC, respectively, were all statistically significant (p < 0.05) in both periods. However, the relationships between RL and TL and NH4+-N, respectively, were not statistically significant for either period. The relationship between RL and NO3−-N was statistically significant for AS but not for BS. The relationship between RL and RF was statistically significant from 2011 to 2018. The temperature dependence of soil respiration was higher in BS than in AS, and the effect of litter removal was 2.55 % and 2.32 % for AS and BS respectively. The results indicate that current global terrestrial models underestimate RL trends for the feedback of global climate change in subtropical forests.