Heart disease and type 2 diabetes represent two of the most pressing global health challenges, with significant impacts onmortality rates and healthcare systems worldwide. The early and accurate prediction of these conditions is critical for effectiveprevention, timely intervention, and efficient management. Traditional diagnostic approaches, while valuable, often lack theprecision and scalability required to address the growing prevalence of these diseases. To overcome these limitations, this studyproposes a machine learning-based predictive system tailored for heart disease and type 2 diabetes, leveraging diverse and extensivedatasets.The proposed framework integrates advanced machine learning techniques, including feature selection, dimensionalityreduction, and cross-validation, to enhance model performance and reliability. A custom ensemble algorithm combining hard andsoft voting classifiers was developed to handle different cardiovascular and diabetes-related datasets. This system achieves robustand high-precision results, with prediction accuracies consistently in the range of 95%.Our framework incorporates essential preprocessing steps such as handling missing data, balancing datasets using SMOTE, andoptimizing features to mitigate biases and improve predictive accuracy. The system is designed to be scalable and adaptable, ensuringcompatibility with diverse real-world healthcare scenarios. By providing personalized risk assessments, it assists healthcareprofessionals and individuals in early detection and informed decision-making.Keywords: Heart failure, Cardiovascular diseases, Type-II Diabetese, Machine Learning.JEL Classification Number: I10, C88, L86.