Black carbon (BC), the residue from burning with insufficient oxygen supply, is assumed to be very stable in the environment. Here we present a simple model for BC movement and decomposition in soils based on the assumption that BC consists of two fractions with different turnover time, and that BC can move in the environment as well as decompose. Decomposition rate was calibrated against laboratory data, whilst a recent field experiment was used to calibrate losses from downward movement through the soil profile. Losses by erosion are still poorly quantified, but mass balance indicates that they may be one of the most important fluxes. The model was able to acceptably predict CO2 production from BC as well as BC left in the soil at the end of the experiment, although BC in the subsoil was underestimated. The model was sensitive to erosion rate (varied ±50%), moisture and temperature response function on a 100-year time scale. The model was not sensitive to the decomposition rate of the stable pool on a 100 year time scale, but it was very sensitive to that on a millennial time scale. Implications and directions for future research are discussed.