SUMMARYOwing to limited wireless network resources, network applications must provide an adaptive qualityguaranteed service to satisfy user requirements. Different applications are associated with different quality of service (QoS) concerns, as well as different QoS control parameters. This work presents an adaptive QoS algorithm by discussing the QoS specifications of three wireless access technologies, i.e. 3G, WiMAX and WiFi. Based on cross-layer and cognition concepts, these environmental parameters are integrated with the sensing of spectral and received signal strength from a cognitive radio paradigm. An adaptive QoS algorithm is then proposed to select the optimal access network for services. Simulation results indicate that the proposed adaptive QoS algorithm outperforms available ones in real-time applications. Compared with traditional algorithms, the proposed algorithm reduces not only the average delay time and jitter for VoIP services to 0.16 s and 0.09 ms, respectively, but also the packet loss ratio for high-definition video streaming by 3.4%.