Here's Yogi Berra's famous answer to your question,
"It's tough to make predictions, especially about the future."
My answer is yes, it is possible sometimes, and it's essential for the advancement of the theoretical sciences, like chemistry. Here's how it SHOULD work most of the time.
First you have a conjecture about how Nature works in some specific situation. Then you frame the conjecture as a testable hypothesis. Then you stick your neck out, and perform an experiment or field study that has the potential to falsify the hypothesis. That's right, you intentionally throw rocks at your pet conjecture.
If the hypothesis survives your experiment and ALL similar experiments by other scientists (within the limits of estimated measurement error), then it becomes a respectable scientific theory. Otherwise you modify it; or chuck it entirely, and go back to the drawing board.
Caveat: Scientific theories--even the best ones--are usually tentative. One well-designed experiment can kill a popular theory, and can trump 1000 'experts'.
I'm old enough to remember when real science was reported in the news, before Post-Normal Science reared its ugly head. But that's another story.
I'm NOT saying that scientists can predict everything. For example, meteorologists cannot make spot-on five-day forecasts consistently. However we do know enough to make reasonably accurate predictions about some things, like the products of some chemical reactions under specified conditions.