Fractal and wavelet methods have been used in this study to analyze the KDP surfaces machined by accurate milling and SPDT (Single Point Diamond Turning) method respectively. Through the 2D wavelet method, the 3D origin machined surfaces were separated into the 3D overlaying roughness surfaces and 3D material structure surfaces. The overlaying roughness surfaces were composed of a large number of length scales superimposed roughness surfaces that are generated from the various vibrations in the machining process. The wavelet method can analyze the information of spatial frequency (vibrations in the machining process) and fractal method can reveal the intrinsic properties of roughness topography. Compared with the conventional methods, the integration of wavelet and fractal is more suitable to characterize the machined crystal KDP surface.