The United States Coast Guard (USCG) conducts the international Ice Patrol (IIP) in the North Atlantic. The primary mission of the IIP is to identify the Limits of All Known Ice (the southeastern, southern and southwestern limits of the iceberg region in the vicinity of the Grand Banks of Newfoundland) and to disseminate this information to mariners. The IIP routinely flies reconnaissance missions during the ice season to help locate the Limits of All Known Ice. This thesis develops an algorithm that, given a set of priorities, determines the optimal routes to fly during these reconnaissance missions. The algorithm relies on partitioning the operation area into squares where the length of each square's side is the IIP's radar or visual identification range. Each square has a reward assigned using IIP priorities which include location of the node, it's proximity to the Limits of All Known ice, whether or not known icebergs are near it, and the time since it was last visited. The algorithm picks the route that conforms to IIP operating procedures with total greatest reward for nodes searched. The algorithm enumerates all routes obeying IIP operational procedures within a few seconds guaranteeing an optimal solution. When compared to actual flights flown by the IIP, routes produced by the algorithm better satisfy USCG defined priorities.