The inactivation model of Salmonella (109 CFU/mL, pool of strains, Salmonella Enteritidis S064, Salmonella Agona IOC1772/11, Salmonella Anatum S156 and Salmonella Senftenberg S087, n=4) submitted to domestic microwave (Electrolux, model 10311LBA106, voltage 127 V, power 1000 W, 60 Hz) through the use of predictive microbiology was assessed. Survival models were constructed as a function of the microwave power level, using add-in GinaFit for Microsoft Excel v1.6. Reductions of 9.22, 9.59, 8.23, and 8.57 log CFU/ mL were observed for the treatments at 20 W (750 s), 40 W (90 s), 60 W (120 s), and 80 W (120 s), respectively, with a maximum temperature rise of 110.2 °C. For all treatments, the models presented good adjustment, once high R² (R² > 0.97) and low RMSE values (RMSE < 0.35) were obtained. For the primary inactivation model, a biphasic profile was initially observed, obtaining a linear log behavior with the increase in power values. Otherwise, the square root model was used for the secondary modeling, resulting in the equation: √kmax = 0.0055 (P + 9.98). For validation of the secondary model, the correlation coefficient R2, the mean square error (RMSE), the bias factor, and the accuracy were used to evaluate the adjustment, being noted good fit of the second model when using the results of with bias, accuracy, RMSE, and R2 values of 0.99, 1.10, 0.04, and 0.98, respectively. Overall, the models demonstrated efficacy to ensure the safety of infant formulas, preventing Salmonella contamination and should be considered considering a practical point of view.