We prove that the set of finite Borel measures on a separable and directionally limited metric space $$(X,\mathtt {d})$$
(
X
,
d
)
is complete with respect to the metric $${\mathbf {d}}_{{\mathcal {A}}}(\mu ,\nu )=\sup _{A\in {\mathcal {A}}}\left| \mu (A)-\nu (A)\right| $$
d
A
(
μ
,
ν
)
=
sup
A
∈
A
μ
(
A
)
-
ν
(
A
)
for all families of Borel sets $${\mathcal {A}}$$
A
that contain every closed ball of X. This allows to prove the existence and uniqueness of the invariant Borel probability measure of certain Markov processes on X. A natural application is a Markov process induced by a random similitude.