Abstract: The pro and con of all classic algorithms for picking up building footprints are analyzed.And then an arithmetic with multi-strategies is put forward to pick up the building footprints from airborne Lidar data.A workflow of this arithmetic is given.Aiming at the deficiencies in settling out data classification,a new measure named neighbor relation is put forward and as a result,points on vegetation surface are effectively separated and eliminated from points on building surface.Aiming at the deficiencies in settling out data classification,r-radius point density is introduced and as a result,the way to choose seed points are improved better and the perform efficiency in segmentation is advanced.