Data stream has been widely used in lots of modern applications such as Social networks and the Internet of things. Aiming at the problem of Top-k dominating query in distributed data stream, a distributed Top-k query algorithm based on Spark Streaming framework is proposed. Based on partitioning, double pruning techniques are implemented on the data. Local and global pruning can significantly reduce the number of candidate sets, reduce the computational overhead and space costs, and improve the query efficiency. Experimental results show that the algorithm has good performance and scalability.