In one of the known data mining models can be used to predict disease stage breast cancer, the artificial neural network model (neural network). this model has a lot of algorithmns, each of which has a different lever of accuracy. algorithm used in this study multilayer perceptron (MLP) and radial basis function (RBF) to predict disease stage breast cancer. both algorihmns are chosen because they are commonly used algorithms. the dara are obained form http://archieve.isc.edu/ml/datasets.html.the data studied is the result of the examination of the 699 people in the ministry of health of wiscoconsin, madison. USA. to implement the algorithm used MLP And RBF NeuralNetwork software SPSS 17.0. The result of the application then dikomparasi using confusion matrix and the ROC curve. based on this study proved that the RBF algorithm has a higher accuracy than MLP.