Osteoporosis is a serious disease characterized by high morbidity and mortality due to atraumatic fractures. In the pathogenesis of osteoporosis, except environment and internal factors, such as hormonal imbalance and genetic background, are also in play. In this study candidate genes for osteoporosis were classified according to metabolic or hormonal pathways, which regulate bone mineral density and bone quality (estrogen, RANKL/RANK/OPG axis, mevalonate, the canonical circuit and genes regulating the vitamin D system). COL1A1 and/or COL1A2 genes, which encode formation of the procollagen 1 molecule, were also studied. Mutations in these genes are well-known causes of the inborn disease ‘osteogenesis imperfecta’. In addition to this, polymorphisms in COL1A1 and/or COL1A2 have been found to be associated with parameters of bone quality in adult subjects. The authors discuss the perspectives for the practical utilization of pharmacogenetics (identification of single candidate genes using PCR) and pharmacogenomics (using genome wide association studies (GWAS) to choose optimal treatment for osteoporosis). Potential predictors of antiresorptive therapy efficacy include the following well established genes: ER, FDPS, Cyp19A1, VDR, Col1A1, and Col1A2, as well as the gene for the canonical (Wnt) pathway. Unfortunately, the positive outcomes seen in most association studies have not been confirmed by other researchers. The controversial results could be explained by the use of different methodological approaches in individual studies (different sample size, homogeneity of investigated groups, ethnic differences, or linkage disequilibrium between genes). The key pitfall of association studies is the low variability (7-10 %) of bone phenotypes associated with the investigated genes. Nevertheless, the identification of new genes and the verification of their association with bone density and/or quality (using both PCR and GWAS), remain a great challenge in the optimal prevention and treatment of osteoporosis.