This paper proposes an algorithm for selecting optimal capacity reservation policy for priority orders in multiple supply plants environment. The proposed approach, first, consolidates the available-to-promise (ATP) from each of the supply plants into cumulated ATP. Then, the capacity reservation policy is considered before the order dispatching is performed. The rate of tardiness for each order and the total revenue generated from order dispatching under various capacity policies are compared, respectively, based on which the optimal capacity reservation policy is selected. Results show that the more the capacity reservation, the higher rate of tardiness for the order with lower value but arriving earlier is. On the contrary, the rate of tardiness for the order with higher value but arriving later is lower. In addition, when the percentage of capacity reservation is increased, the total revenue generated from order dispatching is decreased gradually. This trend is continued until the percentage of capacity reservation is increased to a certain level (turning point). After the turning point, the total revenue is increased in an ascending trend. Thus, the optimal capacity reservation policy can be achieved by finding this turning point. Finally, a numerical example regarding an example with three supply plants to provide capacity for two distinct customer orders is simulated to show the proposed approach. Some managerial implications are provided.