This is simply wrong. In science you can only reject a hypothesis.
Let me illustrate it via an example. Suppose your theory says that there is a positive relationship between two variables. Your null hypothesis is then that the respective parameter in your statistical analysis is zero. If this very hypothesis is rejected in your test (As you do statistics you are in a stochastic realm so there is always a certain likelihood that your test rejects your hypothesis although it is true. Reporting this likelihood is essential when doing empirical work as it should be as small as possible.) you can claim, loosely speaking, that you have indeed proven that there a non-zero relationship between your two variables. What you formally do is nonetheless to reject a hypothesis and not to prove it.