Throughout this master thesis, an evaluation of the video quality degradation due to packet losses in a LTE network is carried out. Indeed, the behavior of the system with respect to packet losses is analyzed thanks to a LTE simulator, identifying at first different LTE scenarios leading to packet losses and analyzing then the packet loss patterns obtained from the simulations for these scenarios. As expected by looking into other studies about packet transmissions over other types of real networks, typical packet loss patterns in a LTE network are bursty. In fact, for low loss rates the bust losses of length 1 (i.e. single packet losses) are the most important loss run lengths. However, when the loss rate increases the burst losses of bigger lengths become more important and their occurrence increases with the increase of the packet loss rate. Therefore, we can conclude that typical packet loss patterns in LTE have a bursty distribution of the losses, i.e. when a specific packet is lost there is high probability that the following packet will be lost as well. Thus, in order to accomplish the quality evaluation, the packet loss patterns identified needed to be modeled. Two mathematical models, which a priori could characterize the burstiness of a specific loss pattern, are identified. Indeed, the validity of both models is verified, i.e. both models are able to model the distribution of burst losses. However, a comparison between them showed that the Logarithmic model is more accurate than the Gilbert model. Finally, the video quality degradation is analyzed taking into account the simulation patterns, the patterns modeled with the Logarithmic model and the patterns modeled with the Gilbert model. Three different input video files, with distinctive characteristics, are used and we could conclude that the packet losses have different effect on the original video clips depending on their features. Moreover, we could observe that the degradation caused by the patterns modeled with the Logarithmic model is closer to the degradation caused by the simulation patterns than the degradation caused by the patterns modeled with the Gilbert model. As a result, the video quality degradation is successfully evaluated and it can be modeled accurately by the Logarithmic model analyzed in this work, enabling to construct degraded video files for use in for example subjective testing.