The aim of the study was to investigate the possible association between the BDKRB2 gene and training-induced improvements in swimming performance in well-trained swimmers. One hundred Polish swimmers (52 men and 48 women, aged 18.1 ± 1.9 years), who competed in national and international competitions at middle- (200 m) and long-distance events (≥400 m), were included in the study. Athletes' genotype and allele distributions were analyzed in comparison to 230 unrelated sedentary subjects, who served as controls, with the χ test. All samples were genotyped for the BDKRB2 -9/+9 polymorphism by polymerase chain reaction. The effects of genotype on swimming performance improvements were analyzed with two-way (3 × 2; genotype × time) analysis of variance with metric age as a covariate. The training period of 1.9 ± 0.4 years had a significant (p < 0.01) effect on swimming performance, both in female and male athletes. Both in female and male athletes, the BDKRB2 gene -9/+9 polymorphism had no significant effect on swimming performance. An interaction effect of BDKRB2 gene -9/+9 polymorphism × time was found for swimming performance only in male athletes. Post hoc analyses showed that swimmers with the +9/+9 BDKRB2 genotype had a greater improvement in swimming performance than swimmers with the -9/+9 polymorphism (p ≤ 0.05). No interaction effects for gender × BDKRB2 gene -9/+9 polymorphism were found for either swimming performance or improvement in swimming performance. These results suggest that the response to long-term exercise training could be modulated by the BDKRB2 gene -9/+9 polymorphism in male athletes. In well-trained swimmers, BDKRB2 gene variation was not found to be an independent determinant of swimming performance.