The effects of material composition, surface chemistry or surface topography on cell attachment (short-term adhesion) have been largely studied on bone-derived cells. However, no statistical demonstration of these effects has been performed until now. With this objective, we quantified the attachment after 24 hours of human osteoblasts on pure titanium, titanium alloy and stainless steel substrates presenting 6 different surface morphologies and 2 different roughness amplitude obtained by sand-blasting, electro-erosion, acid etching, polishing and machine-tooling. The coating by a gold-palladium layer of these surfaces allowed determining the relative effect of the surface roughness and of the surface chemistry. By multiple analysis of variance, we demonstrated that neither material composition nor surface roughness amplitude influenced cell attachment except on sandblasted pure titanium substrates. On the contrary, a high significant influence of the process used to produce the surface was observed meaning that the main influent factor on cell attachment could be either the surface morphology or the surface chemistry induced by the process. As the coating of surfaces by a gold-palladium layer decreased significantly the attachment of cells on the majority of substrates, we concluded that attachment is rather influenced by surface chemistry than by surface topography.