||Today, IEEE 802.11-based Wireless Local Area Networks (WLANs, also known as WiFi) are widely deployed. WLANs are being deployed at everywhere, from home to office, from enterprise to city wide. Traditionally WLAN are operated in infrastructure mode, where all wireless stations are connected to an access point (AP). The range of WLAN is limited by the range of the wireless medium. Wireless mesh networks (WMNs) are the new trends in the wireless world, it extend the reach of WLAN by connecting individual WLANs together wirelessly forming a much larger coverage. This thesis, we describes the design, implementation and evaluation of a hierarchical multi-radio mesh network named Mesh Assisted Channel Assignment (MACA), which normal unmodified 802.11 clients connected to the meshAP via access radio, and the meshAP communicate with each other via mesh radio. Most of the current research efforts are focus on the performance of the mesh radio, while in MACA, we try to utilize the information among mesh nodes to help the access radio channel selection. A key feature of MACA is a new metric to estimate accurately the channel condition under the multi-radio mesh network, the Effective Channel Air Time (ECAT). ECAT can estimate channel conditional correctly when there are traffic in the overlapping channels. Experiments on are being carried out on the MACA implementation to show the effectiveness of the proposed metric and scheme.