Increasingly fierce market competition has encouraged companies to strive for better business, and companies are expected to change their perspective from a product-oriented approach to a customer loyalty-focused approach. Customers are the most valuable asset in running a business. Therefore, loyal customers need to be accurately predicted or to know the level of accuracy in order to help in the decision-making process related to the problem by applying data mining to find patterns hidden in the big data used. The implementation of data mining to predict loyal customers using the Naïve Bayes method based on customer segmentation with RFM modeling can be done well by utilizing sales transaction data. The customer segmentation process is needed to group customers who have similar characteristics by using the RFM model and then produce ten segments. Naïve Bayes is used to forming loyal customer prediction models and the evaluation results of the models that have been tested, namely accuracy of 97.27%, precision of 100%, and recall of 96.98%.