In this study, we modeled the in vitro antibacterial activity of the extracts of green, white and black tea mixtures using the response surface methodology and have characterized its chemical composition. C. sinensis tea mixtures were prepared according to a simplex-centroid design containing 10 experimental assays. The antibacterial activity by agar diffusion method was used to assess the inhibition power of the tea mixtures against a range of seven Gram positive, Gram negative, vegetative cells and sporeforming bacteria. A formulation of tea mixtures was then optimized via the desirability function aiming to maximize the inhibition of different strains and/or serotypes of Salmonella. The best tea formulation was chemically characterized by electrospray ionization mass spectrometry (ESI-MS). The model proposed for in vitro antimicrobial activity of tea extracts on S. Typhimurium was significant (F=10.32, p=0.017). The linear effects of white and black teas were statistically significant (p<0.05) as well as the interaction between green and white and green and black on the inhibition of S. Thyphimurium. A formulation containing 0.60 g of green tea, 1.02 g of white tea, and 0.38 g of black tea maximized the inhibition against S. Typhimurium. Although the optimized tea mixture was effective against all Salmonella strains, inhibition was substantially greater for specific serotypes, with differences observed even among strains of the same serotype (p<0.05). A total of 11 phenolic constituents such as gallic acid, quinic acid, (+)-catechin/(-)-epicatechin, (-)-gallocatechin/(-)-epigalocatechin, coumaroylquinic acid, galloylquinic acid, and caffeoylquinic acid were tentatively identified based on their ESI-MS accurate mass measurements and comparisons with fragmentation profiles. This is the first report on the use of a mixture of C. sinensis teas to produce a water-soluble extract with antimicrobial activity against pathogenic bacteria using design of experiments coupled with MS identification of phenolic compounds that are likely responsible for the observed bioactivity.