A congestion control mechanism which combines compressive sensing and rate control is proposed to solve the congestion problem of wireless sensor network(WSN). For local congestion, by using open-loop hop-by-hop backpressure mechanism, the send rate of the upstream node is adjusted to relief local congestion quickly. For global congestion, the compressive sensing technology is used to reduce redundancy information, thus relieving network congestion. To guarantee the authenticity of information, information is collected in the normal mode after the congestion relieved. Simulation shows that the mechanism achieves better Qos in terms of throughput, latency and energy consumption.