Multi-sensor data association deals with the node information of target units in different environmental backgrounds through characteristic rules, and networking and optimization of sensor networks play a key role in the whole system. At the same time, with the increase of the number of sensors and targets, the complexity of collaborative tracking also increases. In large-scale sensor network data processing, due to the limitation of capacity and processing capability, distributed processing can avoid the impact of single sensor failure on the whole network performance and reduce the communication requirements. Therefore, a multi-sensor network tracking system based on multimedia data is proposed in this paper. The distributed sensor network structure with tree topology is used to transmit multimedia data, and the local fusion center is estimated by parallel filtering. In addition, a distributed multi-sensor network information fusion algorithm (DMSNFM) is proposed, which ensures the correct and consistent interaction of neighboring nodes’ information through channel caching. The simulation results show that compared with the traditional centralized fusion algorithm, DMSNFM can select sensors more reasonably, thus improving the tracking accuracy of maneuvering targets.