Abstract This study classifies local self-government units (local communities) in the Federation of Bosnia and Herzegovina (FBIH) according to socio-economic characteristics by applying the following multivariate methods: principal component analysis (PCA), regression and cluster analysis. The selection of variables was based on literature and adjusted by FBIH specifics, covering the four hypothetical dimensions of regional differentiation: macroeconomic, demographic, infrastructural and socio-cultural indicators. PCA has identified five components: the productivity component, demographic component, component of economic activity potential, spatial component and employment component. Further analysis showed that all identified factors are significant predictors of local communities’ development, measured by the development index. The cluster analysis resulted with four clusters in the FBIH with significant differences in development level. Considering that FBIH municipalities are administrative units of local government and that the classification is based on socio-economic dimensions, identified clusters correspond to the NUTS principles.