An algorithm programmed into the control circuitry of a rechargeable-battery Implantable Medical Device (IMD) is disclosed that can quantitatively forecast and determine the timing of an early replacement indicator (tEOLi) and an IMD End of Life (tEOL). These forecasts and determinations of tEOLi and tEOL occur in accordance with one or more parameters having an effect on rechargeable battery capacity, such as number of charging cycles, charging current, discharge depth, load current, and battery calendar age. The algorithm consults such parameters as stored over the history of the operation of the IMD in a parameter log, and in conjunction with a battery capacity database reflective of the effect of these parameters on battery capacity, determines and forecasts tEOLi and tEOL Such forecasted or determined values may also be used by a shutdown algorithm to suspend therapeutic operation of the IMD.
