The optimal groundwater bioremediation design problem is complex, nonlinear, and computationally expensive. In this paper, an improved Ant Colony Optimization (ACO) algorithm is employed for optimizing a groundwater bioremediation problem, and the BIOPLUMEII model is used to simulate aquifer hydraulics and the bioremediation process. Injection and extraction pumping rates and well locations are treated as decision variables. Optimization results show that the proposed approach performs better than the Genetic Algorithm (GA), Simulated Annealing (SA) and the hybrid SA-GA algorithm, called Parallel Recombinative Simulated Annealing (PRSA), and reduces the computational time of a number of function evaluations compared with the mentioned algorithms. Applying the optimal dynamic pumping strategy in the second stage reduces bioremediation costs by 13:3%.