Bacteremia is an extremely serious infectious disease with high mortality rate and increasing incidence. Early diagnosis and prompt treatment is key to improve the cure rate. In this study, we developed an automated sample pretreatment device based on microfluidic chip and online coupled it with ICP-MS for the simultaneous identification of E. Coli O157:H7 and Salmonella in human blood. This method realized automated sample loading, pathogen capture, elemental labeling, and signal desorption on the microfluidic chip through an automated liquid transfer system and E. Coli O157:H7 and Salmonella were detected simultaneously online by ICP-MS. The linear ranges of 400-80,000 CFU mL-1 for E. coli O157:H7 and Salmonella, as well as a method detection limits of 200 CFU mL-1 for E. Coli O157:H7 and 152 CFU mL-1 for Salmonella were obtained under optimum experimental conditions. The proposed approach realized the automated sample pretreatment and the simultaneous identification of two different kinds of bacteria in human whole blood, avoiding cumbersome sample pretreatment operations and sample contamination in conventional ICP-MS based procedures with elemental labeling strategy for bacteria analysis, demonstrating a strong potential for the clinical applications of bacteremia diagnosis.