Quality of service (QoS) is a key problem in wireless environments where bandwidth is scarce and channel conditions are time varying and sometimes implies highly packet losses. IEEE 802.11b/g/a wireless LAN (WLAN) are the most widely used WLAN standards today, and the IEEE 802.11e QoS enhancement standard exists and introduces QoS support for multimedia applications. This paper presents a class based admission control algorithm for 802.11e based wireless local area networks. The strengths of our admission control is dynamicity and flexibility of the algorithm, which adapts to the situation of the BSS, like global load, number of best effort AC, and position of QSTA by report of QAP, thing that have never been taken together, but each solutions have used a point of sight separately. Thus it achieves higher throughput than other admission control for 802.11 e. The idea which consists of changing parameters [AIFSN, CWmin and CWmax] of best effort flows, for decreasing collisions, is used, so we think that it is an efficient way to protect QoS flows from best effort flows, and to allow reducing then number of collisions. So we increase AIFSN [best effort flow], CWmin and CWmax only at 70% of load of network, to prevent starvations of best effort flows, and so we increase rate utilization of channel. We also use the current rate transmission of QSTAs, according to their positions, instead of the minimum rate transmission used by standard 802.11e, for calculate the load of network and derived the TXOPi necessary for all the stations, with i=1 to number of active stations. The 802.11e standard starves the low priority traffic in case of high load, and leads to higher collision rates, and did not make a good estimate of weight of queues, so there is an unbalance enters the flows with high priorities. A discussion is presented in detail using simulation-based evaluations, with an aim of comparing results of our admission control algorithm, with the 802.11e standard and the F- - HCF algorithm. Results reveal an improvement of the network load and a decrease of the number of collisions.