Poverty that occurs in society one of the problems that experienced by several developing countries, including Indonesia. The main problem that become a concern in various governments, both provincial, district and city, to villages, especially Suka Bhakti Village, Gedungaji Baru  District, Tulang Bawang. As for the ways that have been done to reduce poverty is a social assistance program for the poor. However, in the process of giving social assistance does not have the right criteria, so it is often in implementation the distribution of assistance often occurs in giving errors, after the distribution is done the right criteria as a determining variable for the provision of assistance, giving errors less help. Therefore, the kelurahan hopes to find appropriate technique to provide an accuracy value that the use of these criteria kriteria accurate enough to provide assistance. Thus, the researcher proposes to perform data mining analysis using classification techniques to know and get the classification of society that should receive assistance using six criteria that have been determined by the village. To determine the success of the classification, two algorithms are used: as a comparison between C4.5 and the Naïve Bayes Classifier. From both methods that have been implemented in the data of beneficiary communities with the attributes of the last education of the head of the family, ownership of several assets, the amount of income, building area, type of floor, and type of wall obtained results namely the Naïve Bayes Classifier and C4.5 methods have the same level of accuracy large that is 99.88% with the same low error rate of 0.12% will but the Naïve Bayes Classifier works better than the C4.5 method FPR, Recall, SP and Precision by 100% with a difference of 1% for C4.5 or by 99% performance level from C4.5.