Abstract: Heart is a most important and vital organ in the whole body which performs different functions throughout the day. The term heart attack has become the most familiar disease for doctors in the medical field today. The need to predict the attacks accurately is of major importance as the patients having heart attack are increasing day by day because of many reasons like stress, lack of physical exercise ,bad habits like smoking ,alcohol etc. This paper describes the efficient method of predicting the heart attack using the fusion of both clinical data as well as ECG parameters to increase the accuracy. Probabilistic Neural Network is adopted for training and testing the data. The comparison of results is done for all the three methods and the fusion method is found more efficient.