In　view　of　the　problem　of　too　idealized　modeling　for　magnetic　nanoparticles，a　variety　of　monodisperse　and　polydisperse　magnetic　nanoparticles　(MNPs)　distributions　were　constructed　for　a　proposed　tumor　model．On　this　basis，the　treatment　temperature　for　the　proposed　model　was　predicted　by　solving　the　biological　heat　transfer　equation　after　taking　the　power　consumption　of　MNPs　as　the　input，and　the　therapeutic　effect　differences　due　to　different　dispersed　MNPs　during　magnetic　hyperthermia　were　further　investigated．　In　addition，a　fuzzy　adaptive　proportional-integral-differential　control　system　for　magnetic　hyperthermia　was　constructed　after　considering　a　more　real　distribution　for　polydisperse　MNPs．　This　system　can　accurately　control　the　treatment　temperature　within　a　critical　value　(46°C)　by　temperature　feedback　during　magnetic　hyperthermia，which　solves　the　problem　that　the　local　temperature　of　target　area　is　difficult　to　stabilize　at　the　optimal　safe　treatment　temperature　due　to　the　inhomogeneous　distribution　and　polydispersity　of　MNPs．The　simulation　results　demonstrate　that　the　treatment　effect　can　be　improved　by　increasing　the　aggregation　degree　of　MNPs　size　and　the　uniformity　of　MNPs　distribution　in　the　tumor　area　under　the　same　treatment　conditions．Furthermore，the　control　system　proposed　can　significantly　improve　the　treatment　effect　by　quickly　stabilizing　the　maximum　temperature　to　an　optimal　value，46°C．　©　2024　Huazhong　University　of　Science　and　Technology.　All　rights　reserved.