The liver is a necessary organ of the human body system. Liver disease gives rise to a large number of deaths. Many people suffer from liver disease due to drinking alcohol, drugs, polluted gas, etc. It is important to detect liver disease accurately and timely. Manual detection takes too-much time, expertise, and also prone to error. Deep learning helps a doctor with an automated prediction from an expert system. This article addresses how to predict the condition of the liver in humans through a proposed CNN-based model using blood test results for liver functionality. This will be useful for doctors so they can start their treatment early, and patients have ample time to recover. This article compares ANN, MLP, and RNN techniques with the proposed CNN-based model for predicting liver disease. Results show that the proposed CNN model performs better than other techniques.