Salmonella contamination sources and transmission routes were studied in 5 Belgian poultry slaughterhouses. Samples from the slaughter and cutting line after cleaning and disinfection were collected, as well as neck skin samples and thighs during slaughter of the first flock. In total, 680 swab and water samples were taken from the slaughter line before slaughter. In all slaughterhouses, Salmonella was notwithstanding cleaning and disinfection still isolated from the slaughter line before start of activities. The prevalence of Salmonella in the plucking area was 10.4% (38/365) (hanging area: 5.0%, scalding tank: 5.8%, plucking machine: 17.0%); in the evisceration room, 1.5% (2/138); and in the cutting area, 2.0% (3/149). No Salmonella (0/28) was found in samples from the chilling line. On neck skin samples taken from the various lines, Salmonella prevalence was 16.1% (48/299) after plucking, 16.0% (48/300) after evisceration, 23.3% (70/300) after chilling; on thighs, prevalence was 10.0% (24/240). Nine Salmonella serotypes were identified of which Salmonella Infantis was the most common serovar (53.8%), especially in slaughterhouse A. Two contamination causes were identified; first, although all flocks had an official Salmonella negative status, this was in one case incorrect and led to an enormous contamination of the neck skins of the flock and the slaughterline (i.e., cooling water). Second, molecular typing revealed cross-contamination from flocks slaughtered 1 d before sampling. Salmonella was apparently not always eliminated by the cleaning and disinfection process and able to contaminate the carcasses of the first slaughtered flock. In conclusion, the results of this study provided practical insights for poultry production to further improve their Salmonella control, for example, Salmonella status determination closer to the slaughter date, to adapt cleaning and disinfection protocols especially for critical machinery and better hygienic designed equipment.