In middle-income countries as well as advanced economies of the world, one of the leading health problems is heart disease. Heart disease, otherwise known as cardiovascular disease, is a category of diseases and disorders usually characterized by abnormalities of the heart and the blood vessels, including coronary artery disease, heart failure, arrhythmias, and congenital heart defects. This study is very important as it aims to fill a knowledge gap and tackle the special challenges that primary healthcare systems in areas with few resources face leading to improved health results and more effective healthcare resource usage. In this study, we attempt to assess the efficacy of various machine learning (ML) methods for early heart diseases detection in Primary healthcare. We tested some ML algorithms – logistic regression, random forest, support vector machines (SVMs), K-nearest neighbors, CatBoost and XGBoost (XGB) using medical records from 2022 Primary Healthcare Facilities in Southern Nigeria. Each model showed different strengths in accuracy, precision, recall, and F1 score with logistic regression model achieving an overall accuracy of 85.61%. The Synthetic Minority OverSampling Technique (SMOTE) enabled us to mitigate class imbalance which boosted recall from 0.04 to 0.6016 and also balanced the F1 score from 0.08 to 0.3356; thus accurately identifies heart disease cases while maintaining fewer false negatives. Age (0.570193), daily smokes (0.370494), and blood pressure (0.365896) topped the list of heart risk factors. Blood sugar (0.189080), heart rate (0.096976), BMI (0.052322), and cholesterol (0.047680) also play a part in predicting overall risk.  We recommend adding ML tools into routine healthcare, supported by policies, community outreach, targeted interventions, and continuous research to manage heart disease worldwide