The aim of this study was to investigate the spatial pattern of neonatal and post-neonatal mortality in the city of Goiânia, Central Brazil. Analyses were based on linked birth and death certificates relating to 101,000 in-hospital live births from mothers residing in the city of Goiânia over the 1992-1996 period. Overall neonatal and post-neonatal mortality probabilities were calculated using the linked database. The empirical Bayes method was applied to smooth the estimated rates and minimize random fluctuation. Spatial units of analysis were 65 urban districts, corresponding to the urban planning sectors. The following exploratory spatial analyses were applied: "global" Moran's I statistic, local Moran LISA map, and Gi* local statistics. For both neonatal and post-neonatal mortality there was statistically significant spatial autocorrelation. Results of post-neonatal mortality showed a high-risk cluster located on the outskirts of the city. For the neonatal period, a heterogeneous mortality pattern was found with high-risk districts in all regions, including central areas.