Abstract: Next Generation Wireless Network is Internet Protocol(IP) based that supports any time anywhere service and provide "Always Best Connected" (ABC) state. For real time multimedia service (eg: Video streaming, video conferencing, IPTV and online gaming) high Quality of Service (QoS) should be guaranteed with minimum delay, packet loss and jitters etc. Since NGNs is a ubiquitous wireless communication system seamless mo-bility need to be there to have a continuous service. Different technologies are integrated in to the NGNs and the frequent handover will be there with different technologies and therefore QoS need to maintain. Traffic in the network can be modeled in such a way that it should reduce the delay packet loss and jitters in the NGN environment. Conventional traffic modeling used to model the networks Poisson or Markovian process is inappropriate to model traffic in the networks. The time series Fbm, FGN or FARIMA can be used as traffic model to model the bursty traffic in the network. The FARIMA (p, d, q) is the best fit to model long range dependency as well as the short term dependency traffic in network and predit the future traffic from the present and past traffic. If we are able to predict the traffic in advance resource (mainly bandwidth) can be allo-cate in advance. Network performance like queue length, queue delay, loss probability etc can be anaylsed. Expermental result shows the reduction in buffer size and packet loss so as maintaine QoS in the network.For the queueing analysis and simulation has been done with large deviaton tech-niquie which represent the rare events occur in the most likely way.