The invention discloses a data fusion scheduling algorithm for a wireless sensor network with jointly optimized time delay and power consumption. Firstly, a cluster head node is selected and a fusionrouting trunk tree is constructed to obviously decrease routing nodes and reduce additional power consumption and delay brought by data multi-hop transfer; a data fusion cluster construction algorithmof two-level transmission power is proposed, and, in each of the fusion clusters, the cluster head node is responsible for the data fusion of sensor nodes in the clusters; the compression and dimensionality reduction of redundant data decreases the data transfer volume of the cluster head to reduce the transmission power consumption; and an in-cluster and inter-cluster data fusion scheduling algorithm is proposed to comprehensively consider the factors such as transmission time delay and data deposition, thereby effectively avoiding link conflicts, and ensuring the real-time transmission of perception data.