Soil respiration is a key component of the global carbon cycle, and even small changes in soil respiration rates could result in significant changes in atmospheric CO2 levels. The conversion of tropical forests to rubber plantations in SE Asia is increasingly common, and there is a need to understand the impacts of this land-use change on soil respiration in order to revise CO2 budget calculations. This study focused on the spatial variability of soil respiration along a slope in a natural tropical rainforest and a terraced rubber plantation in Xishuangbanna, Southwest (SW) China. In each land-use type, we inserted 105 collars for soil respiration measurements. Research was conducted over one year in Xishuangbanna during May, June, July and October 2015 (wet season) and January and March 2016 (dry season). The mean annual soil respiration rate was 30% higher in natural forest than in rubber plantation and mean fluxes in the wet and dry season were 15.1 and 9.5 Mg C ha-1 yr-1 in natural forest and 11.7 and 5.7 Mg C ha-1 yr-1 in rubber plantation. Using a linear mixed effects model to assess the effect of changes in soil temperature and moisture on soil respiration, we found that soil temperature was the main driver of variation in soil respiration, explaining 48% of its seasonal variation in rubber plantation and 30% in natural forest. After including soil moisture, the model explained 70% of the variation in soil respiration in natural forest and 76% in rubber plantation. In the natural forest slope position had a significant effect on soil respiration, and soil temperature and soil moisture gradients only partly explained this correlation. In contrast, soil respiration in rubber plantation was not affected by slope position, which may be due to the terrace structure that resulted in more homogeneous environmental conditions along the slope. Further research is needed to determine whether or not these findings hold true at a landscape level.