Reduction in greenhouse gas emissions of existing coal-fired power plants is a necessary action to attain the global reductions committed in the Kyoto Protocol. In the framework of a cap and trade system, we propose a two-stage stochastic mixed-integer linear programming (MILP) approach for the optimal investment timing and operation of a CO2 capture system under uncertainty in the CO2 allowance price. In the MILP, uncertainties are modeled via scenarios that are generated from a set of probability functions obtained using the Geometric Brownian Motion (GBM) approach in conjunction with Monte Carlo sampling. The model takes into account two economic objectives: the expected net profit and the financial risk. We demonstrate the capabilities of the tool presented through a case study based on a coal fired power plant. Our MILP approach can be applied to a wide range of processes and industries that deal with carbon sequestration issues.