Maintaining high levels of physical function is an important aspect of successful ageing. While muscle mass and strength contribute to functional performance in older adults, little is known about the possible genetic basis for the heterogeneity of physical function in older adults and in how older adults respond to exercise. Two genes that have possible roles in determining levels of muscle mass, strength and function in young and older adults are angiotensin-converting enzyme (ACE) and mitochondrial uncoupling protein 2 (UCP2). This study examined whether polymorphisms in these two individual genes were associated with baseline functional performance levels and/or the training-related changes following exercise in previously untrained older adults. Five-eight Caucasian older adults (mean age 69.8 years) with no recent history of resistance training enrolled in a 12 week program of resistance, balance and cardiovascular exercises aimed at improving functional performance. Performance in 6 functional tasks was recorded at baseline and after 12 weeks. Genomic DNA was assayed for the ACE intron 16 insertion/deletion (I/D) and the UCP2 G-866A polymorphism. Baseline differences among genotype groups were tested using analysis of variance. Genotype differences in absolute and relative changes in physical function among the exercisers were tested using a general linear model, adjusting for age and gender. The genotype frequencies for each of the studied polymorphisms conformed to the Hardy-Weinberg equilibrium. The ACE I/D genotype was significantly associated with mean baseline measures of handgrip strength (II 30.9 ± 3.01 v. ID 31.7 ± 1.48 v. DD 29.3 ± 2.18 kg, p < 0.001), 8ft Up and Go time (II 6.45 ± 0.48 v. ID/DD 4.41 ± 0.19 s, p < 0.001) and 6 min walk distance (II 458 ± 28.7 v. ID/DD 546 ± 12.1m, p = 0.008). The UCP2 G-866A genotype was also associated with baseline 8ft Up and Go time (GG 5.45 ± 0.35 v. GA 4.47 ± 0.26 v. AA 3.89 ± 0.71 s, p = 0.045). After 12 weeks of training, a significant difference between UCP2 G-886A genotype groups for change in 8ft Up and Go time was detected (GG -0.68 ± 0.17 v. GA -0.10 ± 0.14 v. AA +0.05 ± 0.31 s, p = 0.023). While several interesting and possibly consistent associations with older adults' baseline functional performance were found for the ACE and UCP2 polymorphisms, we found no strong evidence of genetic associations with exercise responses in this study. The relative equivalence of some of these training-response findings to the literature may have reflected the current study's focus on physical function rather than just strength, the relatively high levels of baseline function for some genotype groups as well as the greater statistical power for detecting baseline differences than the training-related changes.