Cardiovascular diseases have emerged as the leading cause of mortality worldwide in both developed and developing countries in recent decades. Timely identification of cardiac ailments and continuous monitoring by medical professionals can significantly reduce the mortality rate. However, due to the need for greater expertise, time, and discernment, it is not always feasible to accurately detect heart diseases in all cases or provide round-the-clock medical consultation. To address this issue, this research investigation proposes a preliminary design for a heart disease prediction system that operates on a cloud-based platform that utilizes machine learning techniques to detect potential heart diseases. To ensure accurate detection, an efficient machine learning algorithm consequent after a comprehensive investigation of numerous such algorithms must be employed. This paper proposes a method random forest classification; there are various models available that have the potential to enhance classification accuracy. Examples of such models include K-modes clustering, multilayer perceptron (MP), decision tree classifier (DT), Support vector machine (SVM) are used. The application is directed towards a dataset obtained from real-world sources consisting of 85 thousand entities obtained from Kaggle. The decision tree achieved 85.37% accuracy using cross-validation and 85.53% accuracy with no cross-validation, the K-modes achieved 85.87% accuracy through cross-validation and 86.02% accuracy devoid of cross-validation, the Multilayer perception achieved 86.05% accuracy by means of cross-validation and 85.92% accuracy without cross-validation, and the Support vector machine achieved 86.0% accuracy with cross-validation and 85.53% accuracy not using cross-validation. According to the underlying research, the random forest using cross-validation has demonstrated higher performance when compared to other algorithms in terms of accuracy, and the maximum level of accuracy has been obtained at a rate of 87.85%.