This study focused on the effects of organic carbon contents and soil hydraulic conditions on CO2 efflux. Samples were collected at 5 positions (summit, shoulder, backslope, footslope, and toeslope) of the elevation transect affected by erosion and the parent material (loess). Initially, air‐dried soil samples were placed on top of a clay tank, and the samples were wetted by capillary rise to soil saturation, and soil CO2 efflux was measured. Numerical inversions of the measured cumulative capillary rise and evaporation data using the HYDRUS‐1D program were applied to simulate the water regime in the columns and estimate the soil hydraulic parameters. In all cases, the net CO2 efflux (NCER) rapidly increased at the beginning of the wetting. NCER decreased with increasing soil–water content (summit, shoulder, backslope, and loess) or remained relatively stable (footslope and toeslope). The average soil–water content values at the maximal values of NCER (maxNCER) for the summit, shoulder, and footslope were similar. Lower average soil–water contents at maxNCER were simulated for the backslope, toeslope, and loess, which were attributed to the high contents of loess substrate in topsoil samples. The maxNCER measured on topsoils were closely related to the organic carbon contents (R = 0.94) and the maxNCER obtained on all samples correlated with the parameters αRES (R = 0.856) and nRES (R = −0.876) of the soil–water retention curves and saturated hydraulic conductivity (R = 0.856).