The main purpose of this paper is to classify data for patients with heart disease and analysis of models used to predict heart disease patients. Data from the UCI Machine Learning Repository, a Dataset, has 199 samples, including thirteen features, to predict the outcome of cardiovascular disease. The study will start from the collection of data on heart disease. Data preparation and selection is perfect for data mining. To identify people with heart disease. The data were then analyzed using Vertical Hoeffding Decision Tree (VHDT). The result is a technique used to extract data. The experiments showed that data extraction by VHDT and the best results show an accurate is 85.43% and the processing error value is 14.07%. The second root of the smallest expectation is 0.366, suitable for constructing a predictive system for people with heart disease 10-fold cross validation.