For the autonomous flight of Small Unmanned Aerial Vehicle (SUAV) without GPS, a fast Simultaneous Location and Mapping (FastSLAM) based on Improved Particle Swarm Optimization algorithm (IPSO), i.e., IPSO FastSLAM, is presented.An adaptive resampling method is introduced into the proposed algorithm to effectively overcome the degradation of particles and reduce calculation burden on frequency resampling.But resampling process may cause sample depletion, thus diversity measure of particle sets is put forward as an inspiration factor, leading the particles optimization search process after resampling, to ensure diversity particle set optimal.Finally, simulation research is carried out by using the proposed algorithm on SUAV, and result shows that this method can get better performance on estimation accuracy and feasibility.