We exploit liver cancer prediction model using machine learning algorithms based on epidemiological data of over 55 thousand peoples from 2014 to the present. The best performance is an AUC of 0.71. We analyzed model parameters to investigate critical risk factors that contribute the most to prediction.