One important safety issue in automotive industry is the efficient cooling of brake system. This research work aims to introduce an optimized cooling vane geometry to enhance heat removal performance of ventilated brake disks. The novel idea of using airfoil vanes is followed as the basis of this investigation (Nejat et al., 2011, “Heat Transfer Enhancement in Ventilated Brake Disk Using Double Airfoil Vanes,” ASME J. Therm. Sci. Eng. Appl., 3(4), p. 045001). In order to perform the optimization technique efficiently, an integrated shape optimization process is designed. According to the aerodynamic and heat transfer considerations, first an appropriate airfoil is selected as the base profile to be optimized. For the shape modification purpose, a curve parameterization method named class shape transformation (CST) is utilized. The control parameters defined in CST method are then established as the geometrical design variables of an improved territorial particle swarm optimization (TPSO) algorithm. In order to overcome the potential bottleneck of high computational cost associated with the required computational fluid dynamics (CFD)-based function evaluations, TPSO algorithm is coupled with a predictive artificial neural networks (ANN), well trained with an input dataset designed based on the Taguchi method. The obtained profile shows an evident convective heat dissipation improvement accomplished mainly via airflow acceleration over the vanes, avoiding early flow detachment and adjusting the flow separation region at the rear part of the suction sides. The results also reveal the approaches by which such a superior performance is achieved by means of the modified surface curvatures.