The mechanism of rust-preventing action of oils has been discussed and it has been shown that Barnum's theory of water-displacing action of oil seems to be unable to interpret Pilz's experimental conclusion concerning the relation between contact angle and rust-preventive ability of oils. An improved theory has been proposed by which not only Pilz's conclusion can be interpreted but also experimental results of rust-preventive ability of oils can be at least explained as good as Barnum's theory.