Studying and understanding of the surface topography variation are the basis for analyzing tribological problems,and characterization of worn surface is necessary.Fractal geometry offers a more accurate description for surface roughness that topographic surfaces are statistically self-similar and can be quantitatively evaluated by fractal parameters.The change regularity of worn surface topography is one of the most important aspects of running-in study.However,the existing research normally adopts only one friction matching pair to explore the surface topography change,which interrupts the running-in wear process and makes the experimental result lack authenticity and objectivity.In this paper,to investigate the change regularity of surface topography during the real running-in process,a series of running-in tests by changing friction pairs under the same operating conditions are conducted on UMT-II Universal Multifunction Tester.The surface profile data are acquired by MiaoXAM2.5X-50X Ultrahigh Precision Surface 3D Profiler and analyzed using fractal dimension D,scale coefficient C and characteristic roughness Ra *based on root mean square(RMS) method.The characterization effects of the three parameters are discussed and compared.The results obtained show that there exists remarkable fractal feature of surface topography during running-in process,both D and Ra *increase gradually,while C decreases slowly as the wear-in process goes on,and all parameters tend to be stable when the wear process steps into the normal wear process.Ra *illustrates higher sensitivity for rough surface characterization compared with the other two parameters.In addition,the running-in test carried with a set of identical surface properties is more scientific and reasonable than the traditional one.The proposed research further indicates that the fractal method can quantitatively measure the rough surface,which also provides an evidence for running-in process identification and tribology design. Studying and understanding of the surface topography variation are the basis for analyzing tribological problems, and characterization of worn surface is necessary. Fractal geometry offers a more accurate description for surface roughness that topographic surfaces are statistically self-similar and can be quantitatively due as fractal parameters The change regularity of worn surface topography is one of the most important aspects of running-in study. Host, the existing research normally only only one friction matching pair to explore the surface topography change, which interrupts the running-in wear process and makes the experimental result lack authenticity and objectivity. This paper, to investigate the change regularity of surface topography during the real running-in process, a series of running-in tests by changing friction pairs under the same operating conditions are conducted on UMT-II Universal Multifunction Tester. The surface profile data are acquired by MiaoXAM 2.5X-50 X Ultrahigh Precision Surface 3D Profiler and analyzed using fractal dimension D, scale coefficient C and characteristic roughness Ra * based on root mean square (RMS) method. The characterization effects of the three parameters are discussed and compared. The results obtained show that there exists remarkable fractal feature of surface topography during running-in process, both D and Ra * increase gradually, while C decreases slowly as the wear-in process goes on, and all parameters tend to be stable when the wear process steps into the normal wear process .Ra * illustrative higher sensitivity for rough surface characterization compared with the other two parameters. In addition, the running-in test carried with a set of identical surface properties is more scientific and reasonable than the traditional one. The proposed research further indicates that the fractal method can quantitatively measure the rough surface, which also provides an evidence for running-in process identification and tribology design.