I know you can prove this using the splitting lemma but I was thinking about a proof using the snake Lemma.
$0 \longrightarrow N \longrightarrow M \longrightarrow M/N \longrightarrow 0$ and $0 \longrightarrow N \longrightarrow N \oplus M/N \longrightarrow M/N \longrightarrow 0$ are both exact with the obvious homomorphisms and its just as obvious that $N \cong N, M/N \cong M/N$.
So if I found a Homomorphism $\varphi :M \longrightarrow N \oplus M/N$ that commutates i.e. $\varphi(n) = (n, 0+N) \forall n \in N$ and $\varphi(m) = (n, m+N)$ for some $n\in N$, then $\varphi$ must be an isomorphism and $L = \varphi^{-1}(\{(0, m+N)|m\in M\})$.
I'm having some trouble actually describing $\varphi$ im guessing the fact that $M/N$ is free, makes sure $\varphi$ is well defined. So my question is, what is $\varphi$?
In fact, any exact sequence $0 \longrightarrow N \stackrel j\longrightarrow M \stackrel r\longrightarrow F \longrightarrow 0$ of $R$-modules, with $F$ free, splits.
In your case, take $F = M/N$, $j$ the inclusion $n \mapsto n$, and $r$ the canonical projection $m \mapsto m+N$, so that $\ker r = \operatorname{im} j = N$.
If you're still worried about how is $\varphi$ defined, at the end of the proof consider this:
Note that $j$ and $i$ induce a homomorphism $\varphi : N \oplus F \to M$ that maps $(x,y) \in N \oplus F$ to $j(x) + i(y)$. Now, $\varphi$ is an isomorphism: the function $\psi : M \to N \oplus F$ that maps $m \in M$ to $(n_m,r(m))$, where $n_m$ is the unique element in $N$ such that $j(n_m)=m-i(r(m))$ (why such an element exists?), is a homomorphism with the property that $\psi \circ \varphi = \operatorname{id}_{N \oplus F}$ and $\varphi \circ \psi = \operatorname{id}_M$. Thus $M = \operatorname{im}j \oplus \operatorname{im}i$.