The purpose of this paper is to establish the relationship between surface morphology and friction coefficient in the wear process. Different wear stage tests of AISI 52100 ring sliding against AISI 5120 disc were designed and conducted on a rotating setup. The fractal and chaos theories were employed to study the nonlinear features of surface structure and friction signal from spatial and temporal scales. The results showed that 3D surface morphology has fractal nature. The fractal dimension Ds first increased and then stabilized at a maximum and finally decreases dramatically. The multifractal spectrum width Δα presented an contrary evolution trend. The friction coefficient signal has chaotic nature. The standard deviation of distance matrix STD obeyed the evolution rule of a bathtub curve. The correlation value between Ds and STD was − 0.7727, and the correlation value between Δα and STD was 0.7130. The strong correlation between spatial and temporal scales is beneficial to on-line recognition and prediction of wear states in real time.