The cooling system design for air-cooled turbines is a critical issue in modern gas turbine engineering. Advances in the computational fluid dynamics (CFD) technology and optimization methodology are providing new prospects for turbine cooling system design, in the sense that the optimum cooling system of the vanes and blades could be designed automatically by the optimization search coupled with the full three-dimensional conjugate heat transfer (CHT) analysis. An optimization platform for air-cooled turbines, which consists of the genetic algorithm (GA), a mesh generation tool (Coolmesh), and a CHT solver is presented in this paper. The optimization study was aimed at finding the optimum cooling structure for a 2nd stage vane with, simultaneously, an acceptable metal temperature distribution and limited amount of coolant. The vane was installed with an impingement and pin-fin cooling structure. The optimization search involved the design of the critical parameters of the cooling system, including the size of the impingement tube, diameter and distribution of impingement holes, and the size and distribution of the pin-fin near trailing edge. The design optimization was carried out under two engine operating conditions in order to explore the effects of different boundary conditions. A constant pressure drop was assumed within the cooling system during each optimization. To make the problem computationally faster, the simulations were approached for the interior only (solid and coolant). A weighted function of the temperature distribution and coolant mass flow was used as the objective of the single objective genetic algorithm (SOGA). The result showed that the optimal cooling system configuration with considerable cooling performance could be designed through the SOGA optimization without human interference.