As mobile devices become more and more ubiquitous, mobile traffic demands are increasing exponentially. To deal with such challenge, caching is considered as an effective way to reduce redundant transmissions. Cooperative caching allows the cache storage to be used more efficiently and increases cache hit rate to further reduce the downloading latency. Recent researches assume that a mobile user can be served by multiple small cell base stations (SBSs) simultaneously, yet this hypothesis requires costly re-design of the infrastructure’s physical layer. In this paper, we propose a backhaul-based cooperative caching scheme which groups several SBSs into a cluster and the SBSs cooperate through wired interfaces. We first formulate a delay minimization problem for cooperative caching, and then propose a greedy algorithm to obtain the content placement with maximum reduced delay. We also explore the impact on end-to-end latency under different simulation parameters, such as cache size and the skewness of the popularity of requested contents.