The Sanjiang Plain is the largest low-altitudinal swampy plain in China where has been strongly affected by human activities particularly the conversion of marsh to cropland. To evaluate the effect of marsh reclamation on soil respiration, field observations were made at the Sanjiang Mire Wetland Experimental Station (47°35′N, 133°31′E), Chinese Academy of Sciences, located in the eastern part of Heilongjiang Province. Four plots of Carex lasiocarpa marsh, Deyeuxia angustifolia marsh, rice field and dryland were selected in this study. During the experimental periods, measurements were made twice a week in growing season and once a month in the winter. Three replicates were conducted for each plot. Soil respiration rate was measured using the static opaque chamber and gas chromatography system. Results showed that the seasonal changes of soil respiration rates from Carex lasiocarpa marsh and Deyeuxia angustifolia marsh were identical with single-peak. The largest values appeared in summer and the lowest ones appeared in winter. The seasonal variation of soil respiration rate from dryland was the same as the marshland, but the seasonal course of soil respiration rate from rice field changed and the peak values postponed in autumn. The carbon effluxes by soil respiration from Carex lasiocarpa marsh, Deyeuxia angustifolia marsh, dryland and rice field were (3.14±0.45),(4.81±0.68),(2.85±038), (2.18±0.31) tC?hm-2?a-1, respectively. The difference in soil respiration rates between Carex lasiocarpa marsh and Deyeuxia angustifolia marsh was caused by water environment. The lower soil respiration rates from farmland than from marshland was induced by reduced soil organic carbon contents due to marsh reclamation. Temperature and soil moisture/water table were important factors controlling soil respiration. Significant relationships were found between temperature and soil respiration rate from Carex lasiocarpa marsh, Deyeuxia angustifolia marsh and dryland, which could be best described by the exponential equations. Q10 values were 2.1, 2.5 and 1.8, respectively for Carex lasiocarpa marsh, Deyeuxia angustifolia marsh and dryland, and tended to decrease in dryland when compared wtih marshland, which was caused by the decrease in substrate sources in soil after marsh reclamation. The relationships between soil respiration rate and soil moisture/water table were best described by linear or quadratic equations.