Combined with multi-scale analysis ability of wavelet, the wavelet identification of 1/f process is proposed for the characterization of machined surface topography, based on the fractal surface profile obtained by MB function, the fractal dimensions of different wavelet functions and different decomposition scales are calculated,by comparison, the more appropriate wavelet basis function and decomposition scale are selected. The calculation results of the wavelet identification of 1/f process are compared with the box counting method, variance method, R/S analysis method, power spectral density method（PSD）, root mean square method（RMS）, structure function method and the method of equations, and the results show that the wavelet identification of 1/f process is more accurate and more convenient than other algorithms, further shows that the wavelet analysis method can be well applied to the multi-scale fractal analysis on the surface. Finally, the wavelet identification of 1/f process is applied to 3 actual machining surfaces, verified the practicality.