By definition, we have $aNa^{-1}\subseteq N$ for all $a\in G$. This is equivalent to saying $aN\subseteq Na$ for all $a\in G$. (Why?)
Thus, we simply need to show $Na\subseteq aN$ for all $a\in G$.
To this end, take $x\in Na$. Then $x=n_1a$ for some $n_1\in N$, and so $a^{-1}x=a^{-1}n_1a=(an_2a^{-1})^{-1}$. (What should $n_2$ be for this to make sense? Why can I do this?)
Now, because $N$ is normal, $an_2a^{-1}\in N$. But this also implies $(an_2a^{-1})^{-1}\in N$ (Why?) and therefore $a^{-1}x\in N$. Hence, we can write $a^{-1}x=n_3$ for some $n_3\in N$. So $x=an_3$ meaning $x\in aN$.
Hence we have $Na\subseteq aN$, and thus $Na=aN$; or, $aNa^{-1}=N$. However, our choice of $a$ was arbitrary so we have proven our claim for all $a\in G$.