[en] This work addresses the responses of the Atlas Mountain Cedar tree (Cedrus atlantica Manetti) when faced with past and future climactic changes. The ecological study of this species proposes three simulations based on reliable and precise data: - Simulation in 6000 B.P. with the GCM UBRIS-HadCM3M2 climatic data - Simulation in the year 2000 with NEW et al. climactic data (2002) - Simulation in the year 2100 with the ARPEGE Climactic Model data - following the IPCC's scenario A2 These simulations were realized using the CARAIB model, which simulates Carbon stocks in vegetation and Carbon flux between vegetation, soil and atmosphere. In part one, the Moroccan Atlas Mountains, which constitute the principal natural area of the Atlas Cedar tree are described in detail. The different data describing this species' distribution at three periods are summarized: 1) at the present time, 2) at Holocene period and 3) at the end of the Upper Pleniglacial period . The distribution of the species is specified. A map of the Cedar tree's actual distribution in Morocco was drawn from a series of population maps - scale 1/25,000- kept at the Moroccan Ministry of Forestry and Water conservation. So as to improve the sequential study of fossils of Cedar pollen was defined the percentage threshold at which samples containing pollen from the Cedrus atlantica are described as autochtonous or allochthonous. The study of 514 samples of Moroccan Cedar pollen permittted to define this threshold as 1% with minimal risk of error. The last part of the work is focused on modelling. It starts with a description of IBM and CARAIB models used in the three simulations. The data essential for these models are reviewed. Considering the climactic data for these three periods, climate is globally warmer in 6000 BP and in 2100 than today, with the problems of drought becoming more severe in the future. The analysis of the three simulations brings us to two conclusions. Cedrus antlantica is sensitive to climactic change and shows that, at three different periods, its potential distribution areas are quite different. At 6000 BP, the species sought refuge in the high mountains of North Africa. Today's distribution is well simulated in our model. Its good survival in Europe makes it an interesting species for reforestation. In 2100, if IPCC's A2 scenario is validated, modifications in the Cedar's potential and real area of distribution will be considerable. The tree will find little refuge in North Africa. On the other hand, the climate of the land between the Northern Alps and the South of Sweden, associated with a strong concentration of CO2 in the atmosphere may well provide it with an acceptable environment.