As an important kind of clustering protocols in wireless sensor networks (WSNs), LEACH and its variants have been demonstrated to efficiently balance the energy consumption, simplify the routing task and extend the network lifetime. Especially, the multi-hop LEACH type protocols have proved to be effective approaches for high energy efficiency, reliability as well as scalability. However, data transmission in hop-by-hop mode increases the energy consumption and end-to-end delay. In this paper, an improved multi-hop LEACH protocol based on fuzzy logic called IMF-LEACH is proposed, which uses a fuzzy logic controller with residual energy, length of data, and distance to BS as fuzzy descriptors to determine the hop count. Moreover, the hop count is used for each CH to find its optimal next-hop intermediate CH with more residual energy and less number of members. Simulations are conducted in MATLAB to evaluate the performance of the proposed protocol, and the results show that IMF-LEACH maximizes the network lifetime and outperforms its counterparts consistently.