A non-invasive, wearable mechanocardiography (MCG) spectrum assisted cardiac health monitoring and early warning system. Feature points were identified from MCG signals for measuring timing events in cardiac cycles. Physio-markers (i.e. cardiac contractility index) and Left Ventricular Ejection Fraction (LVEF) were calculated. Machine learning algorithms were used to estimate risk potential for heart failure. Early warning will be provided before the onset of the cardiac arrest. The proposed system may become one of the break-through smart devices for mobile detection of heart health and early warning of cardiac attacks at home.