Optimal base station location involves locating multiple base stations within a specific deployment site, while providing an acceptable quality of service to mobiles. Published research has focused on the objective function formulation and local optimization strategies. Local optimization algorithms are not well-suited for the base station location problem due to the inherent nonsmooth, nonconvex and multiple minima that are characteristics of the objective function. A global optimization strategy based on modeling the objective function with a stochastic process is introduced. Comparative results between the new algorithm and several local optimization techniques indicate an improvement of 65% to 90% in determining the global minima and with a decrease in the required number of function evaluations.