Liver cirrhosis is a condition where healthy liver tissue is replaced by scar tissue, causing loss of liver function. Chronic damage from Hepatitis C infection leads to ongoing inflammation, which causes the liver to form scar tissue that disrupts its structure and function, resulting in cirrhosis. This study presents an explainable machine learning framework for predicting liver cirrhosis using clinical data from 1,385 Egyptian Hepatitis C patients monitored over an 18-month treatment period. The dataset comprises 28 real-valued features, with preprocessing steps addressing missing values, discretization, and class imbalance. A two-stage feature selection process—based on Random Forest importance and mutual information—was applied to enhance model efficiency. Multiple ensemble classifiers, including Random Forest, Gradient Boosting, XGBoost, and Extra Trees, were evaluated. Among them, the Extra Trees Classifier achieved the best performance, with 97.01% accuracy, 97.18% precision, and 97.01% recall and F1-score. To ensure model interpretability and clinical transparency, SHapley Additive exPlanations (SHAP) and Local Interpretable Model-agnostic Explanations (LIME) were employed, identifying bilirubin, ALT, and albumin as key predictive features. The proposed framework demonstrates that feature-optimized ensemble learning, combined with explainable AI methods, can support accurate and interpretable cirrhosis diagnosis in Hepatitis C patients, contributing to improved clinical decision-making.