Heat rejection in ground vehicle propulsion systems remains a challenge given variations in powertrain configurations, driving cycles, and ambient conditions as well as space constraints and available power budgets. An optimization strategy is proposed for engine radiator geometry size scaling to minimize the cooling system power consumption while satisfying both the heat removal rate requirement and the radiator dimension size limitation. A finite difference method (FDM) based on a heat exchanger model is introduced and utilized in the optimization design. The optimization technique searches for the best radiator core dimension solution over the design space, subject to different constraints. To validate the proposed heat exchanger model and optimization algorithm, a heavy duty military truck engine cooling system is investigated. For a convoy escort driving cycle, numerical results demonstrate that by increasing the prototype radiator frontal area, the cooling system energy cost can be significantly reduced. The proposed radiator size scaling algorithm offered a methodology of evaluating the tradeoff between physical dimension, heat generation and power. It is a powerful tool that allows optimal deciding the space, cooling and power usage for a given cooling system application.