For a long time the Poisson process has been used to model packet arrivals at Ethernet networks. Recent researchers proved that the Poisson model is inadequate, and self similar processes predict the network traffic better. I was collected traffic samples from one of Mongolian Information Company in Mongolia and their statistical distribution has been analysis study of Wireless network traffic is presented. The characteristics of measured Wireless network traffic on both self simulation and modeling have been investigated. Result have shown Clearly that the self-similar model is more accurate indeed than the Poisson model. In our part we are building the models of self-similar traffic (fractional Gaussian noise), that can be used to simulate various protocols such as Telnet, FTP and WWW packets inter arrival, LAN, WAN, frame relay, ATM, SOHO net.