Abstract: Wireless packet scheduling algorithms, which have been proposed for Quality of Service (QoS) for mobile users in cellular networks, depend on the fundamental idea that the service rate allocated to each user is proportional to a positive weight. Previous work on this problem has typically assumed that the channel conditions between the base station and the users are governed by a stationary stochastic process. However, such stationary assumption is not always reasonable and the study shows that the algorithms based on the stationary assumption, such as a popular algorithm known as M-LWDF，would have extremely poor performance over non-stationary channels. In this paper an opportunistic scheduling model is designed by means of fuzzy logic, which can exploit the time-vary nature of channel conditions to enhance system capacity while satisfying diverse QoS requirements. Simulations illustrate that this design can achieve better performance than previous algorithms under both stationary channel conditions and non-stationary channel conditions.