I have the following problem:
Let $X$ be a Hausdorff Space, and $J \neq \varnothing$, such that $\{K_{j}:j \in J\}$ is a family of compact sets, whereby any finite subset $J_{0} \subseteq J$ shows that $\cap_{j \in J_{0}}K_{j} \neq \varnothing$. Prove that $\cap_{j \in J}K_{j} \neq \varnothing$.
My idea: Let's assume that $\cap_{j \in J}K_{j} = \varnothing$. We now look at $K_{j}^c:=X-K_{j}$. It is clear that $K_{j}^c$ is open as the $K_{j}$ is a compact subset in a Hausdorff Space. We can now say that $X=\cup_{j \in J_{0}}K_{j}^{c}$.
This is where my question comes in: There was no mention made of the compactness of $X$. Is there any way of proving the above without compactness of $X$?
Using compactness, I would say: $\exists N \in \mathbb N: \cup_{j=1}^{N}K_{j}^{c}=X$, and that therefore $\cap_{j=1}^{N}K_{j}^{c}=\varnothing$ which is a contradiction. q.e.d
Your start of the proof will actually show the following:
The proof goes as you started: suppose for a contradiction that $\bigcap_{i \in I} F_i = \emptyset$. Then by define $U_i = X\setminus F_i$, open in $X$, for all $i \in I$, and the empty intersection assumption implies that $\{U_i: i \in I\}$ is a cover of $X$. So by compactness of $X$ we have a finite subcover $X = U_{i_1} \cup \ldots U_{i_n}$ for finitely many $i_1, \ldots , i_n \in I$. It follows (de Morgan) that
$$\bigcap_{j=1}^n F_{i_j} = \bigcap_{j=1}^n (X\setminus U_{i_j}) = X\setminus \bigcup_{j=1}^n U_{i_j} = \emptyset$$
whcih contradicts the fact that the family $F_i, i \in I$ has the FIP. So this contradiction shows that the intersection is non-empty.
Now you're almost where you want to be: we have $X$ merely Hausdorff and a family of $K_j, j \in J$ of compact subsets with the FIP. Fix one $j_0 \in J$ and apply the above lemma to a new $X = K_{j_0}$ (compact) and $F_j = K_j \cap K_{j_0}$ for $j \in J$. This family consists of closed sets because all $K_j$ are closed sets (as $X$ is Hausdorff!) and so these intersections are closed in $K_{j_0}$ by definition of the subspace topology. It still has the FIP, because any finite intersection of them is just a finite intersection of the original family too.
That concludes the proof.