Unit load size is a key factor in an automated guided vehicle based material handling system for a flexible manufacturing system. Highlighting this aspect and its importance at the design stage, this paper presents an integer programming formulation of the problem of finding the optimal unit load size. Using an existing analytical model to decide the number of AGVs required, an algorithm based on branching and implicit enumeration and a heuristic have been developed. Revised computations due to dynamic system conditions such as changes in part mix are also possible with the proposed algorithms. The methodologies have been demonstrated using numerical examples.