Gram-negative　bacteria　had　been　regarded　as　several　important　sources　of　lethal　infection.　Rapid　identification　of　pathogenic　bacteria　is　extremely　important　for　the　diagnosis　and　clinical　treatment　of　diseases.　In　current　study,　three　gram-negative　bacteria,　including　Klebsiella　aerogenes,　Enterobacter　cloacae　and　Escherichia　coli,　were　used　to　access　the　feasibility　of　characterizing　Gram-negative　bacteria　by　surface-enhanced　Raman　Spectroscopy　(SERS).　Bacterial　samples　were　from　Escherichia　coli　isolates　(n=1000),　Klebsiella　aerogenes　isolates　(n=1000)　and　Enterobacter　cloacaeand　isolates　(n=1000).　The　differences　of　three　Gram-negative　bacteria　were　characterized　by　SERS　spectra.　Furthermore,　four　multivariate　statistical　algorithms　based　on　the　combination　of　principal　component　analysis　(or　partial　least　squares)　and　linear　discriminant　analysis　(or　support　vector　machine)　were　used　to　discriminate　the　spectra　of　three　gram-negative　bacteria.