To detect Salmonella gallinarum or Salmonella pullorum in artificially contaminated poultry feed, 9 culture combinations were compared, including 3 preenrichment/enrichment methods (tryptic soy broth plus ferrous sulfate/tetrathionate Hajna, tryptic soy broth plus ferrous sulfate/selenite cystine broth, and Salmosyst) in combination with 3 selective agars (xylose lysine desoxicholate agar added with tergitol 4, EF-18, and Önöz), respectively. Additionally, a single PCR technique was applied combined with 2 different preenrichment media (tryptic soy broth plus ferrous sulfate and Salmosyst). The specificity and positive predictive value were 1 for all methods. There were some differences among Salmonella strains for sensitivity and accuracy in the culture and Salmosyst-PCR methods. The sensitivity and accuracy values were less than 0.60 and 0.64, respectively, whereas the negative predictive values were between 0.12 and 0.23. Two PCR methods did not show any difference in the parameters of performance evaluated. Kappa coefficients showed good agreement between both methods. None of the culture combinations was able to detect S. gallinarum or S. pullorum when the inoculum was less than 3 × 10² cfu/25 g, except the Salmosyst broth method, which could recover S. gallinarum from 3 × 10¹ cfu/25 g onward. Overall, there were differences in the detection limits among the strains and methods used. In general, the 3 selective plating media did not show any significant difference in the parameters of performance studied for each strain. On the other hand, the agreements were slight to fair when culture methods were compared among them and with both PCR methods. The differences in the detection levels that were obtained using these methods and the difficulty in detecting S. gallinarum or S. pullorum in feed represent a potential problem when a poultry feed sample is considered to be negative. It is highly recommended to use at least 2 methods to increase the chances of detecting S. gallinarum or S. pullorum in poultry feed.