A simple methodology is presented which takes input data from wireless accelerometers used in continuous monitoring of large numbers of machines such as pumps, motors, gearboxes, and fans, and outputs a damage accumulation metric which can be used for early warning indication for a broad range of faults. Based on the output, maintenance visits can be made to inspect the machines. Emphasis is on ease of use and broad applicability. It is assumed that the accelerometer data reflects damage occurrence. The approach here is based on time series vibration data analysis that estimates the rate at which damage is accumulated at a given location. This indicator accounts for time-varying symptoms in machines which are often overlooked by traditional vibration diagnostic frequency analysis. As fatigue analysis is the foundation for the damage metric, contribution of repeated load reversal cycles to component damage and the nonlinearity in the relationship between damage and vibration amplitude, are incorporated. A MATLAB code has been developed and validated from simple examples in the literature. The methodology is then applied to a finite element model of a defective shaft-bearing assembly, and in a high pressure pumping field application.