This chapter presents an innovative disease diagnosis method employing a dynamic weighted ensemble approach based on Teaching-Learning-Based Optimization (TLBO). By integrating entropy-based measures and diverse base models like LR, KNN, DT, and SVM, the approach ensures accuracy and robustness. Rigorous evaluations on multiple datasets validate its superiority, offering a powerful solution for healthcare decision-making. Utilizing TLBO within an ensemble learning framework, this approach combines base models including Logistic Regression (LR), K-Nearest Neighbors (KNN), Decision Tree (DT), and Support Vector Machine (SVM). To enhance the accuracy and diversity within the ensemble, entropy-based measures are incorporated, promoting varied predictions and improving overall performance. This inclusion of entropy mitigates overfitting, enhancing ensemble robustness by reducing reliance on any single classifier. The TLBO algorithm continually updates the ensemble’s composition and adjusts weights for accuracy and diversity. Employing a weighted average scheme with dynamically adjusted weights, the approach achieves superior adaptability and accuracy in disease diagnosis. Rigorous evaluation using K-fold cross-validation on datasets like Pima Indian (PID), Statlog Heart Disease (SHD), and Wisconsin-Breast Cancer (WBC) demonstrates the superiority of this model over existing ensemble and non-ensemble methods. Statistical analyses, including the Paired T-test, affirm the distinctiveness of the proposed model compared to baseline approaches. This study presents a streamlined and impact ensemble framework, enhancing clinical decision-making in healthcare applications.