In this work the fractal analysis and the AMT (Angle Measure Technique) method have been considered. These two methods empower us to obtain the estimation of the “complexity” of fracture surfaces structure and to classify the fractured materials depending on the various criteria. Both methods are applied to classify the surfaces, which are simulated with the different values of roughness and noise. In spite of the algorithms similarity and existence of the correlation between fractal dimension and AMT spectra, the results show that application of concrete method depends on the image parameters. For example, if we have a noised image or an image with a non self-similar surface, the better results can be obtained using the AMT method.