Abstract: Wajardikdas 9 years is a compulsory government program that requires the Indonesian people to get a formal education certificate minimally at junior high school level/equivalent. In order to succeed the program, the availability of information relating to the classification of rural/village is absolutely important. Due to this information, the regency/city government can more easily prioritize which rural/village that needs a special attention. There are at least two statistical methods: 1) binary logistic regression and 2) logistic regression. This study compared the results of the classification accuracy between the logistic regression method and tree classification to get a more precise technique used in classifying the rural/village in case of 9 years Wajardikdas program. By using the data testing, it showed that the classification accuracy produced by the method of tree classification was higher than that of logistic regression. Therefore, it could be concluded that tree classification was more effective method to use. This was because the factors that affected the success of Wajardikdas 9 years like the ratio of student/teacher and student/school had been known.