In this paper, we propose an efficient search method for the optimization problem in which discrete variables are present within a limited number of expensive experiments or simulations. Based on the sequential approximate optimization method, a radial basis function (RBF) is employed as a surrogate. A search strategy is used to improve the accuracy of the surrogate and find the global optimum. The search strategy consists of a local search, a global search, and a balanced search to improve the accuracy of the surrogate within the design domain. In order to evaluate the performance of the proposed algorithm, we used unimodal, rugged, and multimodal functions that have been used as discrete variable optimization problems by other researchers. We compared the performance of our algorithm with those of existing algorithms, and showed the effectiveness of the proposed algorithm.