Liver disease is one of the deadliest diseases in the world. Several studies have been conducted to diagnose patients properly but still unknown what method was accurate in predicting liver disease. Data mining is the science that uses past data as a reference to get new knowledge. One of the data mining algorithms is a classification algorithm. Data are obtained from the UCI which consists of 583 records with 11 fields. In this research, comparative Naïve Bayes and C4. 5 algorithms using software algorithms KNAME to know which are the most accurate in predicting liver disease. The results of the second test is known that the algorithm C4. 5 algorithm has the highest accuracy value is 72.845% while the Naïve Bayes algorithm has a value of 63 362% accuracy. Thus C4. 5 algorithm can more accurately predict liver disease.