Based on retrospective modeling of earthquakes from the southern California earthquake catalog, along with previously published evaluations from the New Madrid Seismic Zone, the modified time-to-failure method may be used as an intermediate-term earthquake prediction technique for locating and predicting the size and time of a future mainshock. Modeling previous mainshocks for hypothesis development indicates that the method predicts the actual magnitude of the mainshock to within approximately ±0.5 magnitude units. The error associated with the time-of-failure is approximately ±1.1 years assuming the last precursory event is known. When the last event in the precursory sequence is not known, the predicted magnitude remains similar, but the predicted time will require refinement as additional events are added, with time, to the sequence. The mainshock location can also be identified within a circular region with a radius on the order of tens of kilometers. Criteria are provided for defining acceleration sequences and mainshock locations. The criteria reduce the number of false predictions but also eliminate some mainshocks from our evaluation. Mainshocks as small as magnitude 5.5, occurring between 1980 and 1995, were evaluated from the Southern California earthquake Catalog (SCC). The results were used in association with previous studies to develop a method that can be used for practical (future prediction) applications. The modified time-to-failure method was used to search the SCC for future mainshocks occurring after 17 August 1998. One region satisfied all the criteria and may be modeled by the modified time-to-failure method. The region likely to have a mainshock is a 65-km-radius area centered at 31.43° N, 115.47° W (northern Baja California, Mexico). The predicted magnitude is 6.36, ±0.55, and the predicted time of failure is 1998.565 (7/25/98), ±1.127 years. The addition of future precursory events will allow refinement of the predicted values.