Liver disease is a chronic disease that affects the body and causes death if the patient neglects his treatment. It needs early diagnosis to predict liver disease through clinical tests. For disease prediction, machine learning (ML) models have been used in a traditional manner, where their performance is not suitable for all types of datasets. Thus, the hyperparameter-based machine learning model is proposed to predict liver disease using various ML-supervised algorithms such as K-Nearest Neighbor, Decision Tree, Random Forest, Gradient Boost, and SVM. The proposed model has designed with multiple parameters as per ML algorithms. This model is demonstrated using the Indian Liver Patient Dataset (ILPD) and evaluated with various metric parameters. By employing data balancing techniques, mutual information for feature selection, and Decision Tree-based feature importance analysis, the proposed model demonstrates the potential of machine learning methods to improve early diagnosis and reduce healthcare costs. From our experiments, SVM achieved an accuracy of 98.54% and a F1 score of 99%.