PKH is a program that provides conditional cash transfer to very poor households where they have been determined as PKH participants with certain conditions. Sometimes, this assistance is not in its target. Therefore, it is becoming a problem in this Government empowerment program. This problem arises due to ineffective data verification in the selection of citizens who are eligible to receive PKH assistance. Therefore, it is necessary to analyze PKH for determining the feasibility of a PKH Program problem. Through the results of the PKH analysis, it could be seen whether or not the citizens were feasible to get assistance. Based on this problem, comparative data mining classification methods were used to find out which algorithm is good for predicting citizenship feasibility. There were two algorithms used, namely the C4.5 algorithm and Naïve Bayes. After testing with the two algorithms using RapidMiner tools, the results showed that C4.5 algorithm produced an accuracy value of 91.25% and AUC value of 0.930 with a diagnosis level of Excellent Classification. While the Naïve Bayes algorithm produced an accuracy value of 87.11% and AUC value amounted to 0.923 with a diagnosis level of Excellent Classification. In conclusion, C4.5 algorithm is a good algorithm to be applied to the feasibility of PKH.