With the development of technology and the use of artificial intelligence in everything, especially in health care systems, where health care systems represents great importance in contributing to preserving human lives. Machine learning algorithms play an important role in classifying and predicting diseases.The efficiency of the machine learning algorithms used is measured by accuracy, recall, precision and F1score. The accuracy of machine learning algorithms depends on quality of data set used in training. Therefore, in this paper we present a proposal to predict the severity of liver disease for a data set obtained from the integration of other data sets, namely HCV ILPD, to improve the accuracy of classification and prediction of liver disease.