A cluster-based intersection fingerprinting technique for outdoor location estimation using received signal strength (RSS) is proposed. The performance of the proposed scheme is demonstrated by making comparisons using RSS data from a simulated grid-based urban propagation model, RSS data generated by a network planning tool applied to a rural environment, and RSS data from real city environment. The proposed scheme first uses an optimal clustering scheme to portion the large outside area into different small regions based on the use of RSS deviations from the path loss model. For each region, a fine RSS distribution model is built to provide a good support for further positioning. An improved intersection method is then presented to find the most likely geographical area to further estimate a mobile user's location. A comparison between cluster-based and grid-based environment partitioning is made. The experimental results show that the proposed clustering scheme gives good support for location estimation and the positioning accuracy is significantly improved.