The annotation of SCG signal is very useful in various health care industries, information related to heart problems or cardiac arrest can be fount out successfully annotating various significant points in the SCG signals. Annotating SCG curves accurately is a very difficult task however we predicted few points accurately using machine learning algorithms and with the help of basic preprocessing and dataset creation. The annotation model is classified as a classification problem and we used four binary classifier machine learning models namely logistic regression, Support Vector Classification, Random forest and Gaussian naives based classifier to predict the accurate values of IM and AC points. These points along with the time of appearance of these points we can estimate the cardiac time intervals of a person.The dataset is created by taking the SCG and ECG signals of 20 presumed health volunteers over 3 states and the dataset is available in open-source public repository on physio net “CEBSDB”. The models are then validated using different validation tools like confusion matrix, precision, recall, F1 score, AUC score, log loss and also error rate. We then calculated the cardiac time intervals by finding the difference between time of appearance of these different significant points.