In this study we aimed to detect selection signatures for ten cattle breeds (1,659 animals) representing three major production types (beef, milk, and dual-purpose). By using Illumina BovineSNP50 genotyping assay and FST statistics (standardized and smoothed across genome in 10-SNP sliding window), we have detected selection signal characteristics for separate breeds as well as for breed groups across the production types. This analysis revealed 6 to 12 genome sites per breed spanning the strongest divergent selection signals. Across the whole studied population, the size of these regions ranged from 383 kb to 4.7 Mb, with an average of 1.03 Mb. The regions under selection were distributed on 21 different autosomes. The highest number of divergent selection signals shared by different breeds were found on chromosomes: 6 (n = 17), 14 (n = 9), and 18 (n = 8). Comparison among the production types revealed 5 (milk vs. meat) to 10 (meat vs. dual-purpose) strong selection signals, of which the highest number were detected on chromosomes: 14, 6, and 2. The analysis of gene content of the regions identified genes related to major phenotypic traits differing the breeds. Candidate loci overlapping with genes associated with production (e.g. MSTN, PLAG1), coat color (e.g. MC1R, KIT), disease resistance, general development, fertility, and metabolism regulation were identified and proposed as selection candidates. The presented results provide significant information about possible cause of differences in breed-specific features and are a good complement to current knowledge on genetic differentiation among various cattle breeds.