Due to modern life style, different health diseases are arising day by day. Among that, heart disease makes a human life as very complicated. In Medical field, the prediction of heart disease in early stage has been a challenging one. The main objective of this work is to predict the survival of CHD (Coronary Heart Disease) patients using certain data sets. To overcome ever increasing growth of heart disease, many researchers adopted various kinds of data mining methodologies. Here, the system is designed to discover the condition to find the risk level of patients based on the parameter of symptoms about their health. Several Data Mining Techniques such as Decision Tree (DT), Classification and Regression Tree (CART), Sequential Minimal Optimization (SMO) and Support Vector Machine (SVM) are available to predict heart diseases. The above techniques are employed to predict risk level of patients and provide accuracy as SVM with 84.7%, CART with 85.4%, and SMO with 84.07% and Decision Tree with 89%. Finally, the result showed that DT has a huge potential in predicting risk level of heart disease more accurately.