This paper proposes a novel density-based clustering algorithm, the density-based ordering of clustering objects (DBOCO), for predicting cardiovascular disease through improved data analysis. It integrates data preprocessing techniques, like weighted transform k-means clustering and ensemble feature selection, to enhance prediction accuracy by effectively capturing complex patterns in heart disease datasets. The results indicate that the DBOCO approach outperforms traditional methods in terms of predictive accuracy, providing a reliable framework for early identification and preventive concern in cardiovascular health.