Congestion in a Wireless Sensor Network (WSN) is one of the causes of performance degradation due to severe packet loss that leads to excessive energy consumption. Solutions in WSNs try to avoid and overcome congestion by selecting sensor nodes with sufficient buffer space and adjusting the traffic rate at the source node over the shortest discovered route that usually decreases the End-to-End (ETE) throughput. On-demand routing protocols have the potential to discover the least congested route when it is required. In a WSN, most of the on-demand routing protocols replace the routing metric of the prevalent routing protocol with their proposed routing metric and keep the route discovery mechanism intact, which is not sufficient to increase the performance of the WSN. To address these problems, a novel Congestion-aware and Traffic Load balancing Scheme (CTLS) for routing has been proposed. The CTLS proactively avoids congestion through a novel route discovery mechanism to select the optimum node based on a composite metric. If congestion occurs, CTLS tries to detect it in a timely manner and alleviates it reactively using a novel ripple-based search approach. The simulation results show that the CTLS performs better as compared to the congestion avoidance, detection and alleviation and no congestion control schemes in terms of packet delivery ratio, ETE delay, throughput, and energy consumption per data packet in a resource constraint wireless network. © 2016, Springer Science+Business Media New York.