The　issue　of　coalition　formation　has　been　investigated　from　many　aspects,　but　until　recently　little　attention　has　been　paid　to　overlapping　coalition　formation.　What＇s　more,　an　algorithm　of　multi-task　coalition　parallel　formation　hasn＇t　been　proposed　yet.　In　this　contribution,　we　adopt　the　binary　Particle　Swarm　Optimization　(PSO)　to　address　the　coalition　formation　problem,　taking　overlapping　coalition　and　parallelizability　into　account　simultaneously.　We　develop　an　algorithm　to　parallelize　the　process　of　overlapping　coalition　formation,　and　guarantee　the　precedence　order　of　tasks　at　the　same　time.　As　for　our　algorithm,　a　strategy　is　introduced　to　improve　its　capability　of　global　searching　and　convergent　rate　of　its　solutions.　©　2007　IFAAMAS.