: Regional languages are one of the nation's heritage that must be preserved. According to the results of a survey conducted by UNESCO, 11% of the world's languages are found in Indonesia. The data obtained shows that 9 regional languages in Papua have become extinct and there are more than 50 regional languages that are threatened with extinction. Balinese is one of the regional languages in Indonesia and certainly has a threat of extinction. This study aims to create a stemming algorithm to be able to obtain basic words from affixed words in Balinese. Stemming is one of the most important algorithms needed in information-gathering and text-mining methods. The developed algorithm is an algorithm that initiates research in the fields of text-mining and information retrieval in Balinese. From the results of the research conducted, it is known that the approach using the n-gram analysis technique produces unsatisfactory accuracy. In this study an additional method was used by adding a dictionary to the algorithm designed by matching the basic words with the data in the dictionary. The results of the study show that this method can increase accuracy from 54% to 80% depending on the number of basic words in the dictionary