It's nearly impossible to predict earthquakes so I don't see how this could be feasible at all. Scientists have been trying hard for decades with no luck. It's been one of the most impenetrable natural phenomenons. At most you get decade+ long likelihoods of the number of earthquakes of x strength, across a very large geographic area.

Even in places with a high number of them it's very challenging to give probabilities of when, where, or how often they'll hit.

There's a good reason the predictions for the upcoming US Westcoast major earthquake is always given within the span of centuries.

To harvest energy you need a device on either side of an energy differential. To pull substantial energy from shifting plates you would need something as long, and deep, as those plates. Point devices, ropes being pulled across a slowly separating plate, wouldn't generate much of anything. You would need millions of ropes attached to billions of anchors, all several KM deep within the earth before you could hope to pull any fraction of the energy.