Short‐term changes in arterial stiffness with exercise are proposed to better reflect vascular impairments than resting measures alone and are suggested as a prognostic indicator of cardiovascular risk in older adults. Arterial stiffness indices are reliable at rest, but the time‐course and reliability of postexercise changes in arterial stiffness in older adults are unknown. The precision of postexercise changes in arterial stiffness should be determined prior to their use in large prospective trials. This study assessed the between‐day reliability of the changes in pulse wave velocity (PWV), augmentation index (AIx75) and reflection magnitude (RM) following an exercise bout in older adults. Ten older adults (71 ± 5 years) were tested on three separate days, 7 days apart. PWV, AIx75 and RM were assessed at rest, immediately post and at 20, 40 and 60 min during recovery after moderate‐intensity cycling. Intraclass correlation coefficient (ICC) and reliability coefficient (RC) were used to assess the relative and absolute reliability of arterial stiffness responses. PWV increased, and RM decreased immediately after exercise (P<0·05), and returned to baseline during recovery. AIx decreased during recovery (P<0·001). Resting ICC values were 0·91 (PWV), 0·72 (AIx75) and 0·40 (RM). Reliability of the immediate changes following exercise was high for PWV (ICC:0·87, RC:1·9 m s−1) and moderate for AIx75 (ICC:0·64, RC:7%) and RM (ICC:0·59, RC:9%). Reliability of the postexercise responses was similar to that at rest for all measures of arterial stiffness. These findings indicate that postexercise changes in arterial stiffness indices are reliable in healthy older adults and supports further investigation of the prognostic value of these responses.