Energy efficiency is one of the major challenges in wireless sensor network (WSN) applications. Cluster-based routing is a well-known topology control technique, which is considered as non-deterministic polynomial (NP)-hard. In this work, the problems of energy-efficient clustering approach and routing optimization between cluster heads (CHs) and the sink are studied. We present an energy balanced clustering protocol using binary salp swarm algorithm (ECBS) for heterogeneous WSNs. SALP swarm algorithm (SSA) is a novel meta-heuristic algorithm with strong global search capability. We introduce key factors such as CHs energy, cluster quality, and network coverage into the SSA to find the optimal clustering scheme. In order to further effectively utilize energy and minimize energy expenditure, we propose a method called the two-level gradient forwarding tree determined by a cost function. In this method, the single hop and multi-hop transmission manners for the routing between CHs and the sink are provided. Experimental results indicate that the proposed routing effectively extends the stability period and lifespan of the overall network when compared with other known cluster-based routing protocols.