This article aims to show mineral composition of Mars surface based on TES spectra (Thermal Emission Spectrometer-Mars Global Surveyor), measured in infrared thermal range. It presents how, based on TES data, spectra from selected Martian regions were modelled and interpreted after prior removal of atmospheric influences from the spectra using the Radiative Transfer Algorithm and Deconvolution Algorithm. The spectra from dark area of Cimmeria Terra and light Isidis Planitia were elaborated in cited publications. In the case of light areas ex. Arsia Mons, spectrum of dusty weathered surface of Mars was obtained (also after removal of atmospheric influences) from averaging spectra of dusty regions of Mars. Those aforementioned spectra were used in modelling Martian surface aiming to determine their mineral composition. Deconvolution Algorithm was chosen from the mentioned methods as a tool for the modelling. The spectra described above were used for the Martian surface modelling, such as the Hellas Basin and Martian meteorites SNC (Shergottites, Nakhlites, Chassignites), in order to determine their mineral composition. As a modelling tool one of the following methods of deconvolution algorithm can be chosen. Spectra for the modelling were obtained from the PFS spectrometer (Planetary Fourier Spectrometer) - (Mars Express) and mineralogical composition of basalts from the southern part of Poland were used for this purpose. The method of modelling which was used to determine the mineral composition of Mars and dust can be used in determining mineral composition of selected areas on the Earth from aerial and satellite levels, e.g., soil and vegetation with the use of spectral libraries and spectra of individual plant species.