Here is a generic outline. It may seem obvious, but it is indispensable, and since you have not indicated where your difficulty lies, it is better to start there.
1. Represent your problem as "for all n∈N, P(n)". Here P(n) is the induction statement, also called induction hypothesis in the case of regular (not strong) induction. Note that it depends on n and for each given n it is a proposition, i.e., it is either true or false. In particular, P(n) is not a number.
2. Prove "for all n, P(0), ..., P(n - 1) imply P(n)". This is inductive step. For this, fix an arbitrary n and assume all of P(0), ..., P(n - 1). From here, try to prove P(n).
Consider step 2 in more detail. Let's fix n. If n = 0, then the list P(0), ..., P(n - 1) is empty. This means that P(0) has to be proved from scratch, without any assumptions. When n = 1, you assume P(0) and use it to prove P(1). For n = 2, you assume P(0) and P(1) and prove P(2), etc.
Hint: Here it is enough to have two induction statements P(n - 2) and P(n - 1) as the hypothesis in the inductive step. Note, however, that this does not work for all n. When n = 1, n - 2 is not a natural number. As noted above, to prove P(1) one can only rely on P(0). Thus, for this problem, inductive step can be formulated this way: Prove P(0), then prove P(1) assuming P(0), and for all n >= 2, prove P(n) assuming P(n - 2) and P(n - 1). This last part can be reformulated as follows: for all n >= 0, prove P(n + 2) assuming P(n) and P(n + 1).