To investigate the variation of steel surface with corrosion age, specimens were obtained through rapid corrosion test under neutral salt spray environment and divided into 9 groups according to corrosion age. A PS50 3D profiler was employed to acquire the geometric characteristics and roughness of the test pieces surfaces. Multifractal theory was adopted to calculate the multifractal spectrum of the corroded surfaces, and the characteristic parameters (Δα and Δf) were adopted to study the influences of corrosion age on steel surfaces. Results show that as age grew, the average Δα showed an "increasing (424.34%)—decreasing (26.5%)—increasing (29.7%)" tendency, which was consistent with a fourth power function curve, while the average Δf exhibited a significant decline (53.2%) in a quadratic curve. The increase of corrosion age could result in 1) the style changes of the pits on the surfaces from deep-narrow to wide-shallow and finally a secondary pit, 2) a variation of "significantly increasing—slowing down—stabilizing" the altitude difference of the surface, and 3) a gradually smooth transition between different altitude areas.