Some people have argued that anything fundamentally historical cannot be a science. After all, you can't repeat experiments. How would you repeat Julius Caesar? What lab has built a planet?
Those arguments misunderstand science. Repeatable experiments are important, but the Big Bang is not an experiment. A measurement about the Big Bang is an experiment. If others can repeat and verify my measurement, then I have a repeatable experiment about the Big Bang.
Society agrees that it is possible to become very certain about past events. We hold murder trials, although the jury was not at the murder.
In any case, prediction is much more important than repeatability. Theories about history are scientific if they make predictions that can be tested.
Examples range from the murder mystery and astronomy to geology.
T.H. Huxley wrote a wonderful essay, "On The Method Of Zadig", about how one can make predictions about past events. WARNING: the essay was written in 1880, so you may find it a bit slow.