Abstract: Wireless Access Points (APs) deployed in the same hotspot with a high density will cause channel interference. Considering this issue, also combined with the QoS of terminals in the same Basic Service Set (BSS) and the load balance among different BSS, a three-dimensional discrete Markov chain model based on IEEE802.11e is analyzed. The quantitative relationship of contention window, backoff counter and retransmission number is revised. Besides, a more accurate normalized terminal throughput expression based on QoS is obtained, and the mentioned model is consummated. Secondly, when designing channel allocation, both the QoS of the communication between the terminals within each BSS and AP and the fairness in overall throughput of the terminals within each AP which associated with it and interfered by adjacent AP channel are considered. Also, modeling analysis of channel interference is adopted and the channel allocation issue is ascribed to an optimization issue. Finally, the optimal channel allocation among AP is obtained by genetic algorithm. Numerical analysis results show, compared with Hsum and Channel Assignment based on the Onder of Throughput Reduction CAOTR algorithms, the proposed Channel Assignment based on Fairness and QoS (CAFQ) algorithm based on generic algorithm can minimize the mutual interference among BSS and make a maximum guarantee of the throughput based on QoS within each BSS. Simultaneously, a relatively good fairness in load balance among BSS is obtained.