The　genus　Prorocentrum　includes　many　species　that　are　responsible　for　harmful　algal　blooms　globally,　which　have　serious　economic　impacts　on　marine　aquaculture,　fisheries,　and　tourism,　as　well　as　major　environmental　and　human　health　impacts.　In　this　study,　two　bloom-forming　planktonic　Prorocentrum　species,　P.　cordatum　and　P.　sigmoides,　were　isolated　from　the　coastal　waters　of　China.　Their　morphology　was　examined　in　detail　using　light　microscopy　(LM)　and　scanning　electron　microscopy　(SEM).　The　features　of　their　periflagellar　area　platelet　pattern,　important　taxonomic　criteria,　were　revealed,　and　the　distinctions　in　their　morphology　were　compared　with　other　closely　related　species.　Both　partial　large　subunit　ribosomal　DNA　(LSU　rDNA)　and　internal　transcribed　spacer　region　(ITS1-5.8S-ITS2)　sequence-based　phylogeny　revealed　three　clades　within　planktonic　Prorocentrum,　referred　to　as　P.　micans-like　(where　the　P.　sigmoides　included),　P.　cordatum-like,　and　P.　triestinum-like　groups.　These　groups　showed　consistencies　in　their　thecal　plate　ornamentation　and　periflagellar　area　platelet　pattern　characteristics.　Suitable　habitats　of　P.　cordatum　and　P.　sigmoides　along　the　coast　of　China　were　predicted　with　combined　maximum　entropy　(MaxEnt)　and　random　forest　(RF)　models.　Highly　or　very　highly　suitable　P.　cordatum　habitats　were　mainly　located　in　the　coastal　waters　of　central　to　northern　China.　The　habitat　suit-ability　range　of　P.　sigmoides　was　narrower　than　that　of　P.　cordatum,　and　concentrated　near　the　Yangtze　River　Estuary.　The　main　environmental　drivers　were　identified　by　the　variable　importance　evaluation　procedure　in　BIOMOD　2.　Mean　sea　surface　temperature　(SST)　was　the　main　distribution　driver　of　both　P.　cordatum　and　P.　sigmoides,　and　mean　sea　surface　salinity　(SSS),　distance　from　coastline,　and　phosphorus　concentrations　also　influenced　their　distribution.　Ocean　warming,　eutrophication,　and　re-oligotrophication　could　potentially　expand　the　distribution　range　of　bloom-forming　species.　These　results　showed　that　machine　learning　models　can　play　a　useful　role　in　the　study　of　the　potential　distribution　of　harmful　algal　blooms.