The overall incidence and serotype distribution of non-typhoid Salmonella (NTS) may vary between different geographical localities. To investigate possible regional differences and the effect of demographic factors, we studied 15 865 episodes of laboratory-confirmed NTS infection in Israel. Using Poisson models we found significant variation in the average annual incidence rate of NTS in 15 administrative sub-districts, which was inversely associated with the percent of rural residency (incidence rate ratio 0·75, 95% confidence interval 0·65–0·86, P < 0·001). Variation was also found in the relative incidence of the most prominent serotypes (Enteritidis, Virchow, Typhimurium, Hadar, Infantis), which was affected by rural residency, the percent of non-Jewish population in the sub-district, and the percent of population aged ⩾55 years in the sub-district. Rural residency had a major effect on the epidemiology of salmonellosis in Israel. Future research is required to understand whether decreased incidence in rural areas is an under-detection bias or reflects true differences in NTS illnesses.