Ebola hemorrhagic fever is a disease caused by one of five different Ebola viruses. Four of the strains can cause severe illness in humans and animals. Humans can be infected by other humans if they come in contact with body fluids from an infected person or contaminated objects from infected persons. Humans can also be exposed to the virus, for example, by butchering infected animals. Deadly human Ebola outbreaks have been confirmed in the following countries: Democratic Republic of the Congo (DRC), Gabon, South Sudan, Ivory Coast, Uganda, and Republic of the Congo (ROC), Guinea and Liberia. In this sense, it is of vital importance to analysis the history data and predicts its propagation. More specifically, a model based k-means algorithm to determine the optimal locations of virus delivery is constructed and tested Using Mab-lab programming. By experiment, we find that our model can work well and lead to a relatively accurate prediction, which can help the government forecast the epidemic spread more efficiently.