IP lookup and routing table update affect the speed at which a router forwards packets. This study proposes a new data structure for dynamic router tables used in IP lookup and update, called the Multi-inherited Search Tree (MIST). Partitioning each prefix according to an index value and removing the relationships among prefixes enables performing IP lookup operations efficiently. Because a prefix trie is used as a substructure, memory can be consumed and dynamic router-table operations can be performed efficiently. Experiments using real IPv4 routing databases indicated that the MIST uses memory efficiently and performs lookup, insert, and delete operations effectively.