Heart Attacks are the real reason for death on the planet today, especially in India. The need to anticipate this is a noteworthy need for enhancing the nation's healthcare services. Exact and exact forecast of the coronary illness for the most part relies upon Electrocardiogram (ECG) information and clinical information. These information's must be sustained to a non-direct ailment expectation demonstrate. This non-straight heart work observing module must have the capacity to distinguish arrhythmias, for example, tachycardia, bradycardia, myocardial localized necrosis, atrial, ventricular fibrillation, atrial ventricular ripples and PVC's. In this paper we have built up a productive strategy to procure the clinical and ECG information, in order to group the information in unsupervised way utilizing k-means clustering technique to precisely analyze the heart and anticipate irregularities assuming any. The general procedure can be ordered into three stages. At last we utilize these two information's i.e. ECG and clinical information for grouping by k-means approach for ordering the coronary illness and to anticipate variations from the norm in the heart or it's working by contrasting the class and the real class. We at that point locate the most imperative highlights of the dataset utilizing molecule swarm enhancement and bolster vector machines coupled together and after that apply a similar calculation on the decreased dataset. We can look at the outcomes acquired from the original dataset with the reduced dataset. ©BEIESP.