Cardiac arrest in term of existing research includes only the treatment of adult person and less focus towards the small child and does not include the remote areas. In this paper we try to implement the statistical approach based on SVM (support vector machine) cardiac machine learning dependent model and logistic reversion that is used to design a model to identify the cardiac attack in new born child. The prediction model will act as a service provider to verify the types of attack, ratio of cardiac attack and generate an efficient report to the remote user. The proposed system follow the steps from registration and complete extracted dataset will be analysed in term of level, ration, and type using training and testing procedure. The characteristic stability attained by proposed is about 0.826 and the commonness threshold rate is about 0.842 with a precision of 56. 809%.