Sulfur-containing　metallic　salts　(SCMs)　are　widespread　environmental　pollutants　and　present　varied　toxicity　risks,　posing　potential　threats　to　human　health.　Therefore,　developing　effective　analytical　methods　to　detect　trace　levels　of　SCMs　is　crucial　for　environmental　safety　and　public　health　protection.　In　this　work,　three　iron-based　metal　covalent　organic　frameworks　(Fe-COFs)　nanozymes　with　peroxidase　(POD)-like　activity　were　synthesized　through　the　post-synthetic　metallization　strategy.　Leveraging　their　prominent　POD-like　activity　that　catalyzed　the　oxidation　of　3,3　＇,5,5　＇-tetramethylbenzidine　(TMB)　to　produce　oxidized　TMB　generating　two　characteristic　absorption　peaks,　we　developed　a　six-channel　nanozymes　sensor　array　for　the　identification　and　detection　of　six　SCMs.　Subsequently,　machine　learning　techniques　including　principal　component　analysis　(PCA),　decision　trees　(DT),　random　forests　(RF),　artificial　neural　networks　(ANN),　and　hierarchical　cluster　analysis　(HCA)　were　integrated　to　visualize　array　responses,　enabling　accurate　identification　and　prediction　of　SCMs　in　complex　matrices,　with　a　lower　limit　of　distinction　as　low　as　100　nM.　Furthermore,　the　practical　application　ability　of　the　sensor　array　was　validated　through　the　successful　discrimination　of　binary/ternary　SCMs　mixtures,　unknown　samples　and　actual　samples.　This　study　presents　an　innovative　machine　learning-integrated　multisignal　nanozymes　sensing　platform,　offering　a　novel　avenue　for　the　construction　metal　covalent　organic　frameworks　(MCOFs)-based　nanozymes　sensor　arrays　for　the　intelligent　identification　and　detection　of　SCMs.