Mars is the fourth planet of the Solar System starting from the Sun. Compared to Earth, Mars has an atmosphere that is significantly thinner and undergoes also a seasonal cycle that is more marked. A Martian year lasts about two terrestrial years. One of the particularity of the Martian atmosphere is the continuous presence of an airborne dust layer. SPICAM is an instrument on board the Mars-Express satellite orbiting around Mars since the end of 2003. This instrument is divided in two modules: SPICAM/IR, a infrared spectrometer (1.0-1.7 um) and SPICAM/UV, a spectrometer working in the ultraviolet range (106-317 nm). In this work, we have analysed the SPICAM/UV measurements that have been recorded in nadir viewing and in the spectral interval between 220 and 290 nm. The analysis of this spectral domain allows to study different species and constituents of the Martian atmosphere such as ozone, dust and ice clouds. In the frame of this work, we have developed a method capable of inverting the SPICAM spectra obtained in nadir viewing in order to simultaneously retrieve the integrated quantities of these different species i.e. the ozone total column, and the integrated optical depths of dust and ice clouds. The method is based on a radiative transfer code that allows to take into account the scattering of light by the different molecules and aerosols of the Martian atmosphere. The surface reflectivity is also considered and is retrieved in the cases where no ice clouds are present in the observed scenes, ice clouds reducing the sensitivity in the surface albedo. Therefore, a cloud detection algorithm has been developed and its results were compared with results obtained with other methods and instruments for validation. Our method has been used to invert the data measured by SPICAM/UV over more than four Martian years which allowed to obtain climatologies of the different studied species: the spatial and seasonal distributions of the ozone column, the optical depths of dust and ice clouds and also the surface albedo are presented in this work. These distributions are discussed and compared to those obtained in previous works obtained with other instruments. This work is also a preparation for the future measurement analysis of the NOMAD/UVIS, an instrument that will be on board the ExoMars Trace Gas Orbiter which is planned to be launched in the beginning of 2016. The algorithm developed in this thesis has also been used to assess performances of NOMAD/UVIS such as the detection limits of particular gases (O3, SO2 et NH3) and the values of the signal-to-noise ratio.