Climate changes and global warming are generally associated with the enhanced greenhouse effect, but aerosols can induce a cooling effect and thus regionally mask this warming effect. Unfortunately, the strong variability both in space and in time of the aerosols and thus the difficulty to characterize their global basic properties induce large uncertainties in the predictions of the numerical models. Those uncertainties are as high as the absolute level of the enhanced greenhouse forcing. To solve this problem it is necessary to improve the set of well-calibrated instruments (both in situ and remote sensing) with the ability to measure the changes in stratospheric and tropospheric aerosols amounts and their radiative properties, changes in atmospheric water vapor and temperature distributions, and changes in clouds cover and cloud radiative properties. The quantity used to assess the importance of one compound (greenhouse gases, aerosols) to the variation of the radiative budget of the Earth is the radiative forcing. One of those forcings is the direct aerosol radiative forcing and it depends on the optical depths and the upscatter fraction of the aerosols. Those two parameters depend on the chemical composition and size distribution of the aerosols. Thus the key parameters of this radiative forcing are the chemical composition through its refractive index and the size distribution of the aerosols. This thesis deals with the design and the implementation of one multi-wavelength lidar system at the Jungfraujoch Alpine Research Station (Alt. 3580m asl). This lidar system is a combination of one standard backscatter lidar and one Raman lidar. Its design have been supported by a ray tracing analysis of the receiver part. The laser transmitter is based on a tripled Nd:YAG laser and the backscattered light is collected by one Newtonian telescope for the tropospheric measurements and by one Cassegrain telescope for the future stratospheric measurements. The received wavelengths for each telescope include three elastically scattered wavelengths (355, 532 and 1064nm), two spontaneous Raman signals from nitrogen (387 and 607nm) and one spontaneous Raman signal from the water vapor (408nm). The optical signals received by each of the telescopes are separated spectrally by two filter polychromators. They are build up around a set of beamsplitters and custom design thin band pass filters with high out-of-band rejection. On the visible channel, the adds of a Wollaston prism separates the parallel polarized backscattered signal (532(p)nm) of the perpendicular polarized one (532(c)nm). Photomultiplier tubes perform the detection of the signals for the UV and visible wavelengths and by Si-avalanche photodiodes for the near-infrared signal. The acquisition of the signals is performed by seven transient recorders in analog and in photon counting modes. Within the frame of the EARLINET (European Aerosol Research Lidar Network), hardware and software intercomparisons have been done. The software intercomparison has been divided into the validation of the elastic algorithm and the Raman algorithm. Those intercomparisons of the inversions of the lidar signals have been performed using synthetic data for a number of situations of different complexity. The hardware intercomparison have been achieved with the mobile micro-lidar of the Observatoire Cantonal de Neuchâtel. The present lidar system provides independent aerosol extinction and backscatter profiles, depolarization ratio and water vapor mixing ratio up to the tropopause. Their uncertainties could be smaller than 20% and thus make possible the retrieval of the microphysical aerosol parameters like the volume concentration distribution and the mean and integral parameters of the particle size distribution, (effective radius, total surface-area concentration, total volume concentration and number concentration of particles). This retrieval is performed by one algorithm of the Institute of Mathematic of the University of Postdam based on the hybrid regularization method. The first results of the retrieval of the volume concentration distribution with three backscatter (355, 532 and 1064nm) and one extinction (355nm) profiles has demonstrated promising results. Future upgrades of the system will add ozone concentration and temperature profile up to the stratopause.