Delay Tolerant Networks (DTNs) are sparse and highly dynamic networks where intermittent connectivity is a common event. In these networks, nodes have time-limited connection opportunity to other moving nodes. Therefore, routing is a major challenge in these networks and researchers have tried to design efficient routing algorithms. Important performance metrics of desired routing algorithms are high delivery probability, low energy consumption and low end to end latency. In this paper, we propose an energy-efficient routing algorithm which is based upon the encounter probabilities of nodes. We use these encounter probabilities to derive a utility metric. Moreover, we utilize assumptions and observations of small world experiment to limit active copies and permitted hops of messages. The proposed utility metric and constraints lead us to design an energy-efficient and low overhead routing algorithm for DTN applications. We evaluate the impact of limited buffer capacity and variation of maximum moving speed of nodes on the proposed routing algorithm and compare the results with MaxProp, PROPHET, and SeeR. The first two are well-known DTN routing algorithms and the third is a recently proposed routing scheme for DTN scenarios. Results show that our proposed algorithm reduces the energy consumption of nodes by a factor of 4. It also has a higher delivery probability compared to previously proposed algorithms.