The performance of the data structure has a significant impact on the overall performance of the advance resource reservation in the distributed computing. Because the query and update operations of the B+ tree are of high efficiency, so this paper proposes a B+ tree structure suitable for resource reservation the RRB+ tree. Also, we design and implement the corresponding algorithms of query, insertion and deletion. Different with the B+ tree that insert and delete one key word at a time, the RRB+ tree insert one reservation request and delete one tree node every time. The RRB+ tree is of a higher precision of expression. With the fixed reservation admission control algorithm and the same rate of acceptance, the experimental results show that the RRB+ tree is easier to operate for the complex and changing network environment, and have a higher utilization of storage space.