The objective of the present work was to improve the efficacy and safety of long-term anticoagulation therapy with the use of the methods for molecular-genetic probing of the warfarin metabolism system. The main study group included 156 patients in whom peculiarities of warfarin metabolism were elucidated in a genetic study. The control group was comprised of 150 patients treated in accordance with standard recommendations. The patients remained under the observation during 5 years. The following genes involved in the regulation of the warfarin metabolism system were studied using PCR and RFLP: CYP2C9*2, C430T, CYP2C9*3, A10075C, VKORCI C+1173T. The samples of peripheral venous blood were used for the purpose. The statistical analysis of the data obtained in the study was carried out with a personal computer using the Statistica 6.0 software package. The frequency of alleles and genotypes were determined by direct counting. The deviation of genotype distribution of the studied DNA polymorphisms from the Hardy-Weinberg canonical distribution was estimated using a computer software. It was shown that the patients of the main group were characterized by the prevalence of wild type CYP2C*9 and CYP2C9*3 genes in conjunction with the heterozygous variant of the VKORC1 gene. This finding suggests the predominance of "standard" metabolizers of warfarin. The study demonstrated that the period of induction in the patients of the main group was twice shorter than in controls. Also, the therapeutic doses of warfarin were stably achieved 1.3 times more frequently in the patients who underwent genotyping. The correlation analysis of the relationship between genotyping data and the dosage of therapy allowed determining mean weekly doses of warfarin for different genotype variants. The maximum dose of warfarin was estimated to be 40.7 mg per week in the carriers of the wild type of all the studied genes. It is concluded that the molecular-genetic study of individual peculiarities of warfarin metabolism makes it possible to shorten the period of induction, reduce the risk of hemorrhagic complications (including the fatal ones), and improve the accuracy of prediction of the response to the treatment.