A load-efficient data aggregation strategy in wireless sensor networks(WSNs) is proposed based on compressive sensing(CS). A cluster-based network is constructed, where all nodes send the raw data to their corresponding cluster heads(CHs) and the CHs send the compressed data to the sink through multi-hop path. To realize the load balance among all cluster heads in the networks, an uneven clustering strategy is proposed, and a CS-based load balance(CSLB) strategy is designed. Then a more efficient strategy of Probability-CSLB is proposed to balance the traffic load among all nodes in the network. Finally, a distributed compensation algorithm based on our model. Extensive experiments validate that our scheme can prolong the network lifetime and balance the energy consumption when compared with other schemes.