One of the most important components of a Plug-In Hybrid Electric Vehicle (PHEV) is the battery pack. A large battery pack will allow the car to travel in charge-sustaining mode for a longer period of time, thereby reducing fuel consumption. However, a large pack will result in a high capital and operating cost. In this paper, the optimal number of batteries for the PHEV is determined. The vehicle model was developed using Autonomie and the optimization algorithm was carried out using the actual vehicle model. Due to the large computational time of the problem, the optimization was also executed using an empirical model. Comparisons are provided between the two approaches in terms of accuracy and computational time.