In a computer cluster, using the basic WCF communication architecture, a wide range of DEM data was parallel rendering. According to the task allocation strategy on master node, it was assigned to each server in the cluster; the server used task concurrency model of the WCF; the master node directly distributed tasks to an instance of service which need not put it into the message queue of waiting, to improve the efficiency of parallel rendering of DEM data. Through the different amount of DEM data parallel drawing test on cluster, the rendering acceleration ratio increased with the growing amount of data. For a large amount of data, DEM data parallel processing had important significance.