How do we know? Well, we know relatively: In the layers under our feet, the stuff on top is younger than the stuff on bottom. (Excluding overthrusts, which are obvious).
After all, we couldn't put Layer A ontop of Layer B if Layer B wasn't there in the first place.
So relative ages are pretty simple. (There are, as I noted, mechanisms that can thrust older layers onto of younger ones. But, as I noted, they're really obvious. *really* obvious).
Geologists have long noted (pre-Darwin) that certain fossils are only found in certain layers. Over time, it became a method to quickly identify any given layer. If Fossil X is found only in Layer Y, and you've got a layer with Fossil X, you know what layer it is.
Quite laboriously, geologists built the "geologic column". That is, the history of layering from the beginning up (and I do mean from the part where "it's not all lava anymore"). Layers are named, and sequenced, and it's known what time-frame they were deposited.
All of this was (and still is) constantly checked anytime anyone uses radiometric dating. Radiometric dating is independent of layering, so the fact that multiple independent dating methods all agree with the ages derived from the geologic column is pretty potent evidence.
Now, mind you, radiometric dating isn't without error. Half the papers published on it are about samples that can't be dated, or discussing how a sample was contaiminated, and how to adjust for it. How to recognize it in the field. How to notice it in the lab.
But, it's pretty well known what's reliable when. And, best of all, the methods reach agreement with each other. Which says a lot, as if it were random dates, or skewed dates from the sample, or any of the stuff Creationists like to claim, dating methods would rarely, if ever, agree.
You'll find, if you bother to trace them down, that the Creationist claims of "bad dates" are always examples of samples being dated by methods that cannot be used on them for one reason or another, outright falsehoods, or selective quotes from articles about why a given sample gave a bad date.
Hovind, I believe, is very found of discussing a mollusc that was C-14 dated (Carbon dating is a little different than radiometric dating, as it needs to be calibrated) at an obscenely wrong age. He didn't bother to mention that no one would date molluscs that way because they don't get their carbon from respiration.