Search engine is a program that searches a data in a database based on keywords entered by the user. There are several algorithms that have been used, one of them is Boyer-Moore. However, the use of the method found problems such as slow speed of search and low accuracy of search results when used on a large scale. In this study, is used classification of documents using the Naive Bayes Classifier to overcome these problems. Based on the results of research using 1000 documents, it was found that the speed of searching software with a classified documents is better than the speed of searching software without classified documents. However, the result of accuracy level have the same great value.