In this paper, an optimization technique for a mixed-discrete problem is discussed. The application of a stochastic simulated annealing technique to find the solution of mixed-discrete problem is demonstrated. Although the technique was originally developed to find global optimum, it can be implemented, with simple modification, to a mixed-discrete variable problem. The procedure is applied to optimal designs of stiffened plates. Since the technique requires large numbers of the design constraint evaluation, simplified equations are introduced for the global and the local plate buckling analysis. The results obtained from the study indicate that the mixed-discrete optimization problem of stiffened plate has at least 2 local optima, when it is subject to buckling constraints. The overall procedure of simulated annealing is simple so that it can easily be programmed to find the global optimum of the mixed-discrete problem.