Dissimilar wireless infrastructures such asWiMAX and LTE are fabulously expand in last decade.In this heterogeneous environment of wireless networksand architectures, one of the major concerns is how toallocate network resources efficiently to diverse trafficclasses with different QoS constraints. Further it hasbeen convincingly demonstrated through numeroushigh quality studies that multimedia traffic found inmodern wireless networks exhibits Long-RangeDependence (LRD) and self-similarity, a phenomenonwhich cant’ be captured by traditional traffic modelingbased on simplistic Poisson model.Unlike most existing studies that are primarily based onsimplistic Poisson model and traditional schedulingalgorithms, this research presents an analyticalperformance model for multiple queue systems withself-similar traffic input scheduled by a novel andpromising scheduling mechanism. Our proposed modelis substantiated on G/M/1 queuing system thatconsiders multiple classes of traffic exhibiting longrangedependence and self-similar characteristics. Weanalyze the model on the basis of newly proposedscheduling scheme. We present closed form expressionsof expected waiting times for multiple traffic classes.We develop a finite queue Markov chain for theproposed scheduling scheme. We simulated the variouspatterns of wireless traffic through our own designeddiscrete simulator. The results indicate that ourproposed scheduling algorithm provides preferentialtreatment to real-time applications such as voice andvideo but not to that extent that data applications arestarving for bandwidth and outperforms all schedulingschemes that are available in the market.