The activity of rosemary phenolic extracts against Listeria monocytogenes and Escherichia coli was evaluated in 50% food models of meat, vegetable and dairy products in relation to some factors that can affect MIC (minimal inhibitory concentration) values (inoculum level, proportion of food, temperature) using a new food microdilution method. It was shown that the interactions of meat and milk components with plant extracts reduced the antibacterial effectiveness of rosemary extracts. MIC values for L. monocytogenes were lower than for E. coli in all tested conditions. A lower inoculation level caused a decrease in MIC values for E. coli but an increase in MIC values for L. monocytogenes in control media. In food models, MIC values were higher or equal to MIC values in control media regardless of bacterial type. We showed that the food microdilution method represents a simple, rapid, reproducible and inexpensive method for testing the antimicrobial efficiency of plant extracts in food systems.