In this paper, a distributed traffic-balancing routing algorithm is proposed for multi-sink wireless sensor networks that effectively distributes traffic from sources to sinks. Each node has a gradient field that is used to decide on a neighbor node to reach a sink. The node¡¯s gradient index contains (1) the distance cost from a source to a respective sink, and (2) traffic information from neighboring nodes. The proposed algorithm considers the traffic being faced by surrounding neighbors before forwarding packets to any sink using gradient search for routing and providing a balance between optimal paths and possible congestion on routes toward those sinks. The key objective of this work is to achieve traffic-balancing by detecting congested areas along the route and distributing packets along paths that have idle and underloaded nodes. Extensive simulations conducted to evaluate the performance of the proposed scheme indicate that it effectively reduces the overall packet delay, energy consumption and improves the packet delivery ratio under heavy traffic.