Objectives: To determine the spatial patterns of perinatal mortality in Kocaeli, Turkey using geographic information systems (GIS); to examine whether regional differences exist for the period selected; and whether these differences are linked to regional risk factors. Study design: Ecological research. Methods: Data were obtained from the linked birth-death records data registry maintained by Kocaeli Provincial Health Directorate. Mortality data are added to the geodatabase on a monthly basis. Spatial patterns of mortality rates were determined with GIS by mapping the case differences in the districts, and spatial autocorrelation was used to examine the spatial pattern of mortality rates in the region. Results: Various risk factors contributing to spatial variation of perinatal mortality were revealed in the region. Districts with high mortality rates were shown to be sensitive to these risk factors. The results of this study confirm the direct link between perinatal mortality and poor environmental conditions in the study region. The analyses applied in the study showed that some complex demographic and socio-economic factors should be associated with perinatal mortality rates to identify the geographic patterns of mortality. Conclusions: Implementation of spatial tools within GIS for mortality data showed the efficiency of GIS in perinatal mortality surveillance. This study also demonstrated the capability and utility of GIS to clarify the geographical distribution of perinatal mortality rates in the study area. © 2012 The Royal Society for Public Health.