Poultry products are important in the transmission of zoonotic pathogens, mainly Salmonella. This genus causes millions of foodborne diseases worldwide every year. Cross-contamination by food sources in human cases of salmonellosis and the increase in resistant strains have become important issues. A qualitative and quantitative Salmonella detection method was utilized in a poultry slaughterhouse in São Paulo State, Brazil. We collected 33 samples from different batches of carcasses. Each sample was analyzed at three process points: postbleeding, postdefeathering, and postchilling. A fourth point, retail simulation, was added to simulate retail market storage at 5°C for 72 h. The qualitative methods revealed 100% (33 samples) contamination at postbleeding, 39% (13 samples) contamination at postdefeathering, 58% (19 samples) contamination at postchilling, and 30% (10 samples) contamination at the retail simulation. The quantitative results, determined by the most-probable-number (MPN) technique, ranged from <0.03 to >2,400 MPN/g. We identified 23 Salmonella serovars; the most prevalent were Mbandaka, Senftenberg, and Enteritidis. Resistance to nalidixic acid was significantly more common (P < 0.05) than resistance to other antimicrobial agents. Five multidrug-resistant strains were identified. This study contributes important epidemiological data and demonstrates the need to improve sanitary conditions in slaughterhouses.