Liver disease has become a critical disease showing significant damage on the individuals health with increased rates of mortality across the world. Accurate diagnosis at early stages is useful for preventing severe liver damage. A deep learning-based approach for the prediction of liver diseases using patients medical diagnosis data. By utilizing the features of neural networks, particularly deep convolutional neural networks (CNNs) and deep feed forward neural networks, researcher able to develop models to solve complicated problems related to health disease prediction. In developing model, the dataset is used which includes liver function indicators such as bilirubin levels, enzyme counts, and other patient-specific features. The overall performance of the model is assessed by using standard evaluation metrics such as accuracy, precision and recall. It is observed that deep learning techniques produce better results than conventional machine learning algorithms when the developed models are for prediction of diseases. In this work the deep feed forward neural network and CNN is used to develop the prediction model. The develop model show the accuracy of 97% in predicating the liver disease.