The development of non–destructive techniques is a reality in the field of conservation science. These techniques are usually not so accurate, as the analytical micro–sampling techniques, however, the proper development of soft–computing techniques can improve their accuracy. In this work, we propose a real–time fast acquisition spectroscopic mapping imaging system that operates from the ultraviolet to mid infrared (UV/Vis/nIR/mIR) area of the electromagnetic spectrum and it is supported by a set of soft–computing methods to identify the materials that exist in a stratigraphic structure of paint layers. Particularly, the system acquires spectra in diffuse–reflectance mode, scanning in a Region-Of-Interest (ROI), and having wavelength range from 200 up to 5000 nm. Also, a fuzzy c–means clustering algorithm, i.e., the particular soft–computing algorithm, produces the mapping images. The evaluation of the method was tested on a byzantine painted icon.