At its building, the theoretical new railway line is supposed to be made of perfect straight lines and curves. This track geometry is however gradually damaged and regularly subjected to maintenance operations. The analysis of these track irregularities is a key issue as the dynamic behaviour of the trains is mainly induced by the track geometry. In this con- text, this work is devoted to the development of a stochastic modeling of the track geometry and its identification with experimental measurements. Based on a spatial and statistical de- composition, this model allows the spatial and statistical variability and dependency of the track geometry to be taken into account. Moreover, it allows the generation of realistic track geometries that are representative of a whole railway network. These tracks can be used in any deterministic railway dynamic software to characterize the dynamic behavior of the train.