This is false for at least one definition of limit point.
If we define (as is very common) $x$ is a limit point of $A \subseteq X$ in the space $(X,\mathcal{T})$ iff for every $O \in \mathcal{T}$: if $x \in O$, then $O \cap (A \setminus \{x\}) \neq \emptyset$.
Another definition (which I learnt as adherent point): $x$ is an adherent point of $A \subseteq X$ in the space $(X,\mathcal{T})$ iff for every $O \in \mathcal{T}$: if $x \in O$, then $O \cap A \neq \emptyset$.
Then we have that $$\overline{A} = \{x: x \text{ adherent point of } A \}\text{.}$$
And $$\{x: x \text{ limit point of } A \} \subseteq \overline{A}\text{,}$$ without there being equality. The set $\{x: x \text{ limit point of } A \} $ is also denoted $A'$, the so-called derived set of $A$. $A$ could have isolated points, i.e. points $x \in A$ such that $\exists O \in \mathcal{T}: O \cap A = \{x\}$, and then for all $A$: $\overline{A} = A' \cup \{x: x \text{ isolated in } A\}$.
E.g. $A =\mathbb{Z}$ in the reals in the usual topology has $A' =\emptyset$, i.e. no limit points, but it equals its set of adherence points, so is closed.