Heart disease diagnosis has become a difficult task in the field of medicine. This diagnosis depends on a thorough and accurate study of the patient’s clinical tests data on the health history of an individual. The tremendous improvement in the field of machine learning aim at developing intelligent automated systems which helps the medical practitioners in predicting as well as making decisions about the disease. Such an automated system for medical diagnosis would enhance timely medical care followed by proper subsequent treatment thereby resulting in significant life saving. Incorporating the techniques of classification in these intelligent systems achieve at accurate diagnosis. Neural Networks has emerged as an important method of classification. Multi-layer Perceptron Neural Network with Back-propagation has been employed as the training algorithm in this work. This paper proposes a diagnostic system for predicting heart disease. For diagnosis of heart disease 14 significant attributes are used in proposed system as per the medical literature. The results tabulated evidently prove that the designed diagnostic system is capable of predicting the risk level of heart disease effectively when compared to other approaches.