The variations of fiber bundle surface microstructure have direct influence on the material performance, especially the friction and wear properties. Therefore, fiber bundle is the smallest evaluation unit of C-f/SiC composite surface. However, due to the anisotropy and inhomogeneity of C-f/SiC composite, it is difficult to evaluate the surface characteristics. Researchers think that two-dimensional evaluation is not suitable for the composites surface assessment any more because of its complex composition and varied surface structure. In this paper, a novel method is introduced for the evaluation of 3D needled C-f/SiC composite fiber bundle surface. On the level of C-f/SiC composite fiber bundle surface, two-dimensional evaluation method is adopted, with which the fiber bundle surface quality can be quantitatively evaluated by the two-dimensional surface roughness Ra. As long as the extracted surface profiles averagely distributed on C-f/SiC composite fiber bundle surface, with appropriate sampling length and sampling number, the mean value of Ra can estimate the whole surface roughness, thus reflecting the roughness degree of surface accurately. This study not only benefits the detection of 3D needled C-f/SiC composite fiber bundle surface quality, and lays a foundation on the evaluation of material functional features in further. And it corresponds to the convenient application in engineering practice. (C) 2015 Elsevier B.V. All rights reserved.