CO2 emissions from soils were measured under 3 land-use types at the adjacent plots of forest plantation, grassland, and cropland from January 2005 to December 2006. Mean soil CO2 efflux rates measured during the 2-year study varied from 59 to 527 mg CO2/m2.h in forest plantation, 37 to 498 mg CO2/m2.h in grassland, and 32 to 397 mg CO2/m2.h in cropland. Soil respiration in the 3 types of land-use showed a similar seasonal pattern in variation during both years, in which the single-peaked curve occurred in early summer and the minimum in winter. In particular, the date of maximum soil CO2 efflux rate in cropland occurred about 30 days earlier than in forest and grassland in both 2005 and 2006. The relationship of soil respiration rate (R) with soil temperature (T ) and soil moisture (W ) fitted well to the equation R = β0eβ1TW β2 (a, b, c were constants) than other univariate models which consider soil water content or soil temperature alone. Soil temperature and soil moisture together explained 69–92% of the temporal variation in soil respiration in the 3 land-use types. Temperature sensitivity of soil respiration (Q10) was affected positively by soil moisture of top 0.1 m layer and negatively by soil temperature at 0.05 m depth. The relationship between Q10 values and soil temperature (T ) or soil moisture (W ) indicated that a 1°C increase in soil temperature at 0.05 m depth will reduce the Q10 value by 0.07, 0.05, and 0.06 in forest, grassland, and cropland, respectively. Similarly, a 1% decrease in soil moisture of the top 0.1 m layer will reduce the Q10 value by 0.10, 0.09, and 0.11 in forest, grassland, and cropland.