Refrigerant　substitution　is　an　important　topic　in　developing　air-source　heat　pumps　(ASHPs).　The　new　refrigerants　were　selected　based　on　their　environmental　impact　indices,　thermodynamic　performance,　safety,　and　economy.　However,　their　effects　on　frosting　have　always　been　neglected.　To　quantify　the　variations　in　the　frosting–defrosting　performance　of　ASHPs　using　different　refrigerants,　twenty-one　ASHPs　from　nine　manufacturers　were　selected　and　tested　at　2　℃/1　℃.　The　most　commonly　used　refrigerants,　R410A　and　R32,　were　employed　in　this　study.　Variations　in　frosting　operating　time　(Tf)　and　optimal　defrosting　initiation　time　(Topt)　were　investigated　for　different　values　of　the　characteristic　indice　for　configuration　and　operation　(CICO).　The　results　show　that　the　frosting–defrosting　performance　of　ASHPs　is　significantly　affected　by　the　type　of　refrigerant.　The　frosting　rate　of　ASHPs　using　R32　was　31%–48%　faster　than　that　of　R410A　for　a　similar　heating　capacity.　The　Topt　of　ASHPs　using　R32　should　be　advanced　by　29%–35%.　Based　on　the　experimental　results,　two　prediction　models　for　the　Tf　and　Topt　of　ASHP　using　R410A　and　R32　were　established.　They　predicted　the　Tf　and　Topt　values　by　calculating　the　CICO　values,　which　can　be　used　to　inform　the　design　of　frost-suppressing　mechanisms　and　the　implementation　of　accurate　defrosting　procedures　for　ASHPs.　©　2024　Elsevier　Ltd