Let k be a field. Then k[[x]] is an Euclidean domain.
Proof
We define v:k[[x]]∖{0}→Z≥0 by setting v(f) to be the degree of the dominant term of f (which we will refer to as the order of f).
Suppose a,b∈k[[x]] with b non-zero. If b is a unit, then we can divide by b without remainder, which suffices for Euclidean valuation. Suppose otherwise.
Let r∈k[[x]] be the polynomial containing all monomials of a whose degree in x is less than the order of b. Note that, by construction, v(r)<v(b).
We want to find q∈k[[x]] such that a=bq+r. To do this, we will define a series of approximations that converge to the desired q.
Let q0=0, and let a0=a−r. Observe that the invariant a0=a−r−q0b holds at this step.
Suppose qn−1 and an−1 are defined, and the required invariant holds for them. Then we define their successors as follows. If an−1=0 then an=0 and qn=qn−1. Otherwise, let sxm be the dominant monomial of an−1. Since k is a field and m≥v(b), there exists a monomial s′xm′ such that the dominant term of s′xm′b is sxm. Let qn=qn−1+s′xm′ and let an=an−1−s′xm′b.
Since an−1=a−r−qn−1b, we observe that
an=a−r−qn−1b−s′xm′b=a−r−(qn−1+s′xm′)b,
which validates our invariant. Also note that, as long as both an−1 and an are non-zero, v(an−1)<v(an), as subtracting s′xm′b removes the dominant monomial of an−1.
Define a metric d(f,g) on k[[x]] by d(f,g)=2−v(f−g) if f=g and d(f,g)=0 if f=g. By construction, d(a−r,bqn)→0 as n→0. By completeness of k[[x]], this limit exists.
Taking q to be the limit of q0,q1,…, we now can write a=bq+r. Since r is either 0 or has valuation v(r) less than b,v is indeed a Euclidean valuation on k[[x]]. Thus, k[[x]] is a Euclidean domain.