Abstract : Geometrical and energetically models are widely used in outdoor and indoor noise predictions. Several approaches have been considered, amongst them, one can cited the ray and beam-tracing methods, the image-source method, the radiosity method... Another approach is based on the use of sound particles to model the sound propagation in complex environments. Similarly to the ray-tracing methods, the sound particles concept allows to model many acoustics phenomena like absorption, transmission, reflection and diffusion on surfaces, scattering by fitting objects, meteorological effects... Two approaches can be followed: a probabilistic one, considering that the energy that is carried by the particle is constant and using Monte-Carlo methods for the modelling of acoustics phenomena; a deterministic one, considering that the energy of sound particles is modified according to the acoustics phenomena. In this paper, a numerical code (SPPS), using both approaches is presented and many applications are shown. In addition, several comparisons with experimental data and other methods are also proposed.