We must assume $G$ is nontrivial, of course, though the problem does not mention it (I guess it is implied by the existence of a nontrivial normal subgroup...).
If $G/Z(G)$ is abelian, then $G$ is nilpotent (of class at most $2$). If $G$ is finite, then it is the product of its $p$-parts, and the problem is reduced to the problem for $p$-groups, which is solved here for nilpotent groups of any class.
But here’s an argument that does not require finiteness of $G$ or knowing about the nilpotency: note that because $G/Z(G)$ is abelian, then $[G,G]\subseteq Z(G)$.
Let $h\neq e$ be an element of $H$. If $h$ is central in $G$, there is nothing left to do. If $h$ is not central in $G$, let $g\in G$ be an element such that $gh\neq hg$. Then
$$e\neq [h,g] = h^{-1}g^{-1}hg = h^{-1}(g^{-1}hg)\in H.$$
But we also have $[h,g]\in [G,G]\subseteq Z(G)$, thus exhibiting a non-trivial element of $H\cap Z(G)$.