Abstract: Compared with the traditional Internet, in-networking caching is one of the most distinguishable features in named data networking (NDN). In NDN, a node caches every passing data packet as a default model. The caching scheme generates a large number of redundant data in in-networking. As a consequence, the networking cache resource is wasted seriously. To solve the problem, a caching scheme based on node classification (BNC) is proposed firstly in this paper. Based on different node positions, the nodes that data packet passes through are divided into two types: “edge” type and “core” type. When data packet passes through the “core” type nodes, by considering location and data popularity distribution at different nodes, it is cached in a node which is beneficial to other nodes. When the data packet passes through the “edge” nodes, a node is selected through data popularity to be beneficial to the client. The simulation results show that the proposed scheme can efficiently improve the in-network hit ratio and reduce the delay and hops of getting the data packet.