Abstract: Abstract: Enterobacterial Repetitive Intergenic Consensus (ERIC) primarily exists in enterobacteria. The sequence length is about 124-127 bp, which contain about 44 bp highly conservative core sequences in its center. ERIC-PCR was widely applied to research the intestinal microbial community composition and diversity. However, we rarely saw domestic and foreign related reports about the ERIC-PCR method which was adopted to analyze the microbial community compositon in raw milk. ERIC-PCR molecular biotechnology was used to research the microbial community, and DNA fingerprints on microbes in raw milk in North China and in Central South China was established in this paper. First, the bacteria DNA extraction method in raw milk had to be built. This paper utilized the CTAB extraction method. Secondly, we needed to determine the optimal ERIC-PCR reaction condition. In order to obtain a distinct and high quality fingerprint, influencing factors on the ERIC-PCR system for amplification was optimized, which included annealing temperature, Mg2+ concentration, primer concentration and Taq DNA polymerase enzymatic activity. Randomly selecting four DNA samples, each condition set four gradients. The best level of all the factors were that the optimum annealing temperature was 52℃, the optimum Mg2+ concentration was 2mmol/L, the optimum primer concentration was 0.24μmol/L and the optimum Taq DNA polymerase enzymatic activity was 2.5U. Thirdly, a DNA fingerprint on microbe in raw milk in North China and in Central South China had to be established. ERIC-PCR product was conducted through electrophoresis in 1.7% agarose gel, then the electrophoresis product was analyzed by a Gel Imaging System. We utilized Quantity One 4.6.2 software to analyze the ERIC-PCR DNA fingerprint. Diversity analysis of the fingerprint adopted the Shannon-Wiener index. The computational formula is H= -∑Pi lnPi (Pi=Ni/N, Ni represents the i ERIC band area under the peak of one sample and N is on behalf of all the ERIC band area under the peak of the sample.) Similarity analysis of the fingerprint was calculated automatically by the software according to a Dice coefficient, then phylogenetic tree diagram was drawn by using the weighted average of the distance (UPGMA) analytical method. By comparing DNA fingerprints of two regions, we could draw a conclusion that the microbial community structure of different raw milk samples from near areas were highly similar and had some dominant microflora respectively, while the different raw milk samples from distinct areas had lower similarity. This study indicated that the establishment of a specific DNA fingerprint in different regions could provide scientific proof for the monitoring of quality and safety in raw milk.