Abstract: In this paper, we propose a service discovery strategy based on dynamic correlation-aware method. Firstly, the nodes of the network are divided into multiple logical regions, according to the spatial correlation of nodes and the temporal correlation of sensor data. Secondly, considering the sensor node's processing ability, cache capacity and communication range and so on, the dictionary node can be selected reasonably. And then the parameter matching service provides nodes based on the distance of the logical region, thus completing the whole services discovery process. Results show that the reliability of service discovery can be enhanced by our proposed mechanism.