The geographical patterns of genetic variation in domestic species are the outcome of spatially explicit demographic events. Technological advances have made the obtaining of spatial information easier and different techniques have been developed to analyse spatial patterns of genetic variation. Relatively weak geographic trends found in sheep populations suggested the use of more powerful approaches, such as spatial principle component analysis. The aim of the study was to assess if application of spatial approach could reveal structures and patterns of genetic variation in indigenous Eastern Adriatic sheep breeds and contribute to our current knowledge about their genetic differentiation. We found south-east to north-west cline as the global structure using spatial principal component analysis, which outperformed the principal component analysis in this study and complemented understanding of variability of the investigated breeds reported in other multivariate and model-based clustering methods applied on this microsatellite data in previous research. Kernel density estimation suggested the Lika pramenka sheep belongs to a separate patch, not recognized in the spatial effects of the spatial principal component analysis. This potential structure should be further investigated.