In this paper, we propose a space-time dynamic model for describing the temporal evolution of greenhouse gas concentration in the atmosphere. We use this dynamic model to develop an optimal control strategy for reduction of atmospheric pollutants. We prove the existence of optimal policies subject to control constraints. Further, we present necessary conditions of optimality using which one can determine such policies. A convergence theorem for computation of the optimal policies is also presented. Simulation results illustrate removal of greenhouse gas using the optimal policies.