Cardiovascular disease (CVD) is considered as one of the global health threats, where early detection is key to effective treatment and preventing complications. In the past, there were manual procedures to treat and identify the disease which gradually was assisted by advanced medical tools. The application of ML techniques along with the traditional ways has shown drastic changes in the rate at which diseases were diagnosed and also cured. This paper explores the potential of ML to improve CVD detection accuracy. We review existing research in the field of ML-based CVD detection using some of the main techniques SVM, Random Forest, Logistic Regression, Gradient Boosting, and Deep Neural Network (DNN), which were identified to be efficient as per the survey. To compare these models, we utilize UCI dataset via Kaggle and present a comprehensive evaluation based on metrics such as accuracy, precision, recall, F1-score, and AUC-ROC. Our investigational outcomes illustrate that DNN outperforms the other models, achieving an accuracy of 93.32%, precision of 92.18%, recall of 94.43%, F1-score of 93.29, and an AUC-ROC value of 0.845. These findings suggest that DNN holds promise as a powerful tool for early CVD detection.