That is not quite correct. Some axioms systems are categorical. This means that there is only one model up to isomorphism. For instance, take the collection of all statements about the Natural numbers (using the standard model of them under the first order Peano Axioms). This collection forms an axiom system for the Natural numbers and all true statements are provable in this system. Indeed, every true statement is an axiom.

The problem with doing this is that the collection of all true statements about the Natural numbers is not recursively enumerable. There is no effective (algorithmic) procedure for determining when a given statement is an axiom or not. This means that under this system we can't find all true statements, even theoretically, using a computer because of the non recursively enumerable nature of the axiom system.

The first order system provides a nice recursively enumerable set of axioms. But these axioms are not powerful enough to deduce all true statements about the Natural numbers. The second order axioms are powerful enough but are not recursively enumerable.

Recursively enumerable was the goal of Hilbert and others because they wanted to reduce mathematics to an algorithmic or mechanical process of verification. That isn't possible and this is the main reason why Roger Penrose (and me) think that computers will never be able to do mathematically what humans can do.