Peak-Force Quantitative Nano Mechanic (QNM) AFM is an edge-cutting scanning probe microscope and can be used to quantitatively investigate mechanic properties of materials in the nanoscale. With force mapping technique, the different phases could be distinguished according to their different mechanic features. For some macromolecular complex systems, however, the phase contrast of their force mapping images is fuzzy, resulting in that the phase identification might very depend on personal judgment of the researchers and some tiny microstructure could not be detected. In order to resolve this problem, we developed a new methodology for objective and accurate phase identification, basing on the deconvolution of modulus distribution of the force mapping images. In consideration with the fact that different phase exhibits different elastic modulus with normal distribution, the modulus distribution curve the image of the macromolecular complex system was fitted into several Gauss peaks; and accordingly how many different phases in the complex phases can be identified. The peak position and peak area reflect the characteristics modulus and relative fraction of corresponding phase. According to the deconvolution result, the modulus mapping images was recolored so that the size, shape and space distribution of identified phases can be demonstrated visually. This new method was used to analyze modulus mapping images of several complex systems, including functionalized SSBR/performance resin, isoprene rubber under stretching deformation, and carbon nanotube reinforced NR nanocomposites. Some hidden but valuable microstructural information were disclosed, facilitating deep understanding structure-property relationship.