It's simply a matter of definition, nothing more. You need to fully understand inverse functions and the conditions under which a function has an inverse. e^x takes any number and outputs a positive real number (I'm assuming you're restricting your attention to real numbers) and it's also one to one and onto the positive reals, so it has an inverse function that takes any positive number and returns some real number.
So you have some number x and you exponentiate to get e^x. Now, every positive number y is the result of e raised to some power (it's not a proof, but looking at the graph of e^x will probably convince you of this). This power is called ln(y).
A verbal description would be that ln(y) is the power you need to raise e to to get y. Then e^ln(y) is e raised to the power you need to raise e to to get y. In other words, e^ln(y)=y. When it's stated that way, it's not so mysterious.