Random forest plays an essential role in various classification and prediction problems, especially in the field of medical diagnosis. Diagnostic prediction of cirrhosis is extremely important for early prevention. Based on clinical data, this study constructs a multi-view depth-bounded random forest (MVDBRF) model, combining clinical data with decision tree variables in a random forest. MVDBRF can achieve the early prediction of cirrhosis. We used MVDBRF to process 583 data samples, with a determined tree depth of 5 and 131 decision trees in the forest, resulting in a prediction accuracy of 95.62%. Compared with previous studies, MVDBRF is powerful for prediction and can effectively predict the correlation with disease regarding a specific dataset. It is highly promising as a computer-aided predictive diagnostic method for biomedical research.