Mobile-to-Mobile (M2M) communication is a challenge, but fast growing since it has the potential to enhance the quality of data service with low energy consumption. This dual mobility generates non-stationarity properties of the propagation channel, which requires precise modeling, for simulation network and system design. This thesis aims for characterization and modeling of M2M channels based on stochastic approaches. Some common M2M aspects like the non-stationarity, Doppler and correlation properties were also analyzed. Stochastic channel models were proposed for Body to Body (B2B) and Vehicle to Pedestrian (V2P) channels based on measurements. The non-stationary fading behavior were observed as a function of mutual body orientation and the relative Tx-Rx distance. A general model for the auto-correlation function of the fading channel was derived assuming a typical M2M scenario in the presence of moving scatterers; and a simplify Doppler spectrum model was also developed for indoor M2M scenarios. Both models allow to reproduce with good accuracy the correlation and Doppler properties of different M2M scenarios including line-of-sight (LoS) and non-line of sight (NLoS) environments. Finally, implementation techniques able to reproduce the fading characteristics of M2M channels including the Doppler and correlation characteristics were proposed and successfully compared to the measurements.