Customer loyalty is an important factor to the survival of a company. The problem facting today is related to the descreasing number of customer loyalty that happened to PT MNC Play. If this is allowed   to continue it is not impossible this will endanger the continuity of the company s business.Keep in mind the factors that cause customers to have loyal and disloyal status, data mining classification techniques can be used to classify loyal and disloyal customers. Many data mining classification algorithms can be used, so it needs to be comparative to know the accuracy of each algorithm, the algorithem used is C4.5 and Na ve Bayes. The data used were 28,899MNC Play customer of Semarang City.The results of the classification   process were evaluated using cross validation, confusion matrix, ROC Curve to find the more accurate data mining classification algorithm to determine loyal and disloyal customers.Keywords: Classification, Customers Loyalty, Na ve Bayes, C4.5.