Abstract: Ubiquitous in-network caching is one of the key aspects of Content-centric Networking (CCN) which has recently received widespread research interest. At present, researchers usually use Leave Copy Everywhere (LCE) strategy as the cache decision strategy by default. Experimental investigation shows that LCE apparently is not a good choice because it does not effectively distinguish the importance of nodes on content delivery path. By learning from the advantages of some existing basic cache decision strategies, a new cache decision strategy which called Content Gradually Tend to Important Node (CGTIN), is proposed in this study. CGTIN is a cache decision strategy who tries its best to push more popular contents to more important nodes and extend the survival time of these more popular contents. In order to test the performance of CGTIN, programs of 4-level binary tree topology are developed by MATLAB. Simulation results show that while compared to LCE, both the SSHP and AHD of CGTIN has an apparent decline while compared to MCD, the SSHP of CGTIN has an apparent degradation and the AHD of CGTIN is exact same when not considering the class of CCN content.