This paper proposes an automatic base station planning approach. This approach does not only try to assess usual objectives such as radio coverage, but further implements a quality of service (QoS) constraint. This criterion is here defined as the mean available bandwidth per user. Its computation takes the medium access control (MAC) layer behavior, the multiple bit rates of IEEE 802.11b and the coverage area of each access point (AP) into account. A Markov chain is used to evaluate the available bandwidth of each cell independently. This is sensible because a non-overlapping criterion between cells is used during the planning phase to avoid interference between adjacent cells. The planning process estimates an optimal number of APs and their placement minimizing an aggregate criterion using a Tabu meta-heuristic. The AP's coverage is estimated with a powerful multi-resolution radio propagation simulator previously described. This paper presents results for a QoS oriented planning process providing a predefined minimum peruser throughput. The example provided herein applies for 100 users distributed over a 12600 m 2 building floor