The　performance　evaluation　method　based　on　data　envelopment　analysis　(DEA)　is　one　of　the　important　tools　to　measure　the　competitiveness　and　productivity　of　enterprises.　However,　the　input　and　output　of　enterprises　may　contain　negative　data　and　the　essence　of　DEA　is　an　iterative　optimization　model,　resulting　in　a　low　applicability　of　the　DEA-based　performance　evaluation　method　in　the　real　word,　especially　for　the　dilemma　of　evaluating　enterprise　performance　within　a　limited　time　for　new　enterprises.　Therefore,　this　study　firstly　develops　a　DEA　model　that　can　handle　negative　data　for　enterprise　performance　evaluation,　and　then　further　establishes　a　new　method　base　on　the　extended　belief　rule-base　(EBRB)　model　for　enterprise　performance　online　evaluation.　A　case　study　about　35　Chinese　state-owned　enterprises　are　conducted　to　verify　the　effectiveness　of　the　proposed　enterprise　performance　online　evaluation　method.　Experimental　results　showed　that　the　proposed　method　has　capable　of　evaluating　enterprise　performance　with　accurate　efficiency　values　better　than　some　existing　performance　evaluation　methods,　and　its　computation　time　is　significantly　less　than　the　DEA-based　performance　evaluation　method,　which　guarantee　that　the　proposed　enterprise　performance　online　evaluation　method　can　serve　as　a　reference　for　the　promotion　of　enterprise　productivity　and　sustainable　economic　development.