Conotoxins(CTXs)　can　specifically　act　on　multiple　ion　channels,　which　are　crucial　for　the　development　of　neurobiology　and　novel　targeted　drug　development.　At　present,　＞10,000　kinds　of　CTXs　have　been　sequenced,　it　would　be　extremely　laborious　to　conduct　experiments　for　each.　μ-CTX　KIIIA　is　a　type　of　substance　that　can　selectively　recognize　voltage-gated　sodium　ion　channels.　This　article　constructs　four　derivatives　of　KIIIA　and　predicts　their　3D　structures;　afterwards,　their　molecular　orbital　arrangements　and　physicochemical　properties　were　calculated　using　DFT;　then,　predicted　their　toxicokinetic　parameters　such　as　absorption,　distribution,　metabolism,　excretion　(ADME)　and　toxicity　(T)　through　Machine　Learning　(ML);　finally,　molecular　docking　and　molecular　dynamics　are　used　to　investigate　the　interaction　modes　and　binding　affinity.　The　results　indicate　that　the　toxicity　of　KIIIA　and　its　derivatives　(KIIIA-1　-KIIIA-4)　to　the　human　body　is　mainly　concentrated　in　the　liver　and　respiratory　tract.　Among　four　derivatives,　KIIIA-2　(5　Ser　→　Arg)　has　better　toxicokinetics　properties　and　its　binding　energy　to　Nav1.2　is　−65.32　kcal/mol,　which　is　higher　than　that　of　wild　type(−32.13　kcal/mol).　This　study　indicate　that　computational　toxicology　can　facilitate　the　druggability　research　of　CTXs,　and　KIIIA-2　can　be　developed　as　a　potential　antiepileptic　drug.　©　2025　Elsevier　B.V.