Over the past years, heart failure (HF) or congestive heart failure (CHF) has been classified as a growing and widespread epidemic worldwide that significantly impacts morbidity and mortality, especially in the aging groups. This research paper adopted a publicly available clinical dataset collected from 299 patients with HF. The dataset consists of 12 parameters, including age, anemia, creatinine phosphokinase, diabetes, ejection fraction, high blood pressure, platelets, serum creatinine, serum sodium, sex, smoking, and follow-up time potentially contributing to mortality. Several research papers identify the crucial factors influencing the patients' mortality using the dataset. Here, we apply data curation to the dataset, ensuring that the dataset is unbiased, and principal component analysis and machine learning models to identify the factors that contribute to the crucial parameters that contribute to the mortality of the patients. Here, we also investigate and compare the classification accuracy of different machine learning models, including tree model, linear discriminant model, quadratic discriminant model, logistic model, Naive Bayes, support vector machine, nearest-neighbour model, ensemble and kernel model. The ensemble model with the bagged tree has the highest cross-validation classification accuracy of 96.4% and requires only three parameters, including platelets, creatinine phosphokinase, and the follow-up period.