Activated carbon injection (ACI) is an effective mercury control technology demonstrated in both short-term and long-term full-scale tests. The effectiveness of mercury capture by activated carbon depends on the mercury speciation, total mercury concentration, flue gas composition, method of capture, and activated carbon properties, such as pore size, type of carbon impregnation, and surface area, etc. It is also desired that an ACI system be designed to produce good mixing between the activated carbon and the flue gas. In recent years, General Electric Energy has conducted both short-term and long-term tests in large-scale coal-fired boilers for ACI mercury capture demonstration. The programs consisted of (1) combustion optimization to improve natural mercury capture by fly ash, (2) computational fluid dynamics (CFD) modeling of activated carbon injection to design ACI lances, (3) a short-term test to select the activated carbon type, and (4) a long-term test to evaluate the mercury capture performance. This paper presents the CFD modeling for an ACI demonstration in Sundance Station Unit 5. The CFD model developed describes the film mass transport, pore diffusion, and carbon surface adsorption and desorption phenomena for the prediction of the mercury capture rate. The model was applied to evaluate the lance design and to calculate the mercury capture rate. The test data were also presented for comparison with the model results.