(1)　Background:　Our　study　aims　to　explore　the　impact　of　abusive　management　and　self-efficacy　on　corporate　performance　in　the　context　of　artificial　intelligence-based　human-machine　interaction　technology　in　enterprise　performance　evaluation.　(2)　Methods:　Surveys　were　distributed　to　578　participants　in　selected　international　companies　in　Turkey,　Taiwan,　Japan,　and　China.　To　reduce　uncertainty　and　errors,　the　surveys　were　rigorously　evaluated　and　did　not　show　a　normal　distribution,　as　it　was　determined　that　85　participants　did　not　consciously　fill　out　the　questionnaires,　and　the　questionnaires　from　the　remaining　493　participants　were　used.　By　using　the　evaluation　model　of　employee　satisfaction　based　on　a　back　propagation　(BP)　neural　network,　we　explored　the　manifestation　and　impact　of　abusive　management　and　self-efficacy.　Using　the　listed　real　estate　businesses　as　an　example,　we　proposed　a　deep　learning　BP　neural　network-based　employee　job　satisfaction　evaluation　model　and　a　human-machine　technology-based　employee　performance　evaluation　system　under　situational　perception,　according　to　the　design　requirements　of　human-machine　interaction.　(3)　Results:　The　results　show　that　the　human-machine　interface　can　log　in　according　to　the　correct　verbal　instructions　of　the　employees.　In　terms　of　age　and　education　level　variables,　employees＇　perceptions　of　leaders＇　abusive　management　and　self-efficacy　are　significantly　different　from　their　job　performances,　respectively　(p　＜　0.01).　(4)　Conclusions:　artificial　intelligence　(AI)-based　human-machine　interaction　technology,　malicious　management,　and　self-efficacy　directly　affect　enterprise　performance　and　employee　satisfaction.