Background: Self-selected exercise intensity (SSEI) promotes higher affect (pleasure) during exercise, but its reliability is still unclear. Purpose: to evaluate the test-retest reliability and the minimum detectable change of the perceptive and physiological responses in two sessions of SSEI on treadmill in elderly women. Method: Twenty elderly women (ages 65.3 ± 4.2 years) performed two 20-min laboratory-based treadmill aerobic exercise sessions with self-selected intensity. During the sessions, %VO2max, %HRreserve, affect, and rating of perceived exertion (RPE) were recorded. Reliability was calculated using the intraclass correction coefficient (ICC) and Bland-Altman plots. The minimum detectable change (MDC) was also calculated. Results: ICC values were 0.98 for % VO2max, 0.83 for % HRreserve, 0.85 for affect, and 0.80 for RPE. No differences were observed in mean values between sessions for all the variables. MDCs were lower than 0.7% for VO2max, 11.7% for HRreserve, 0.7 for affect, and 0.8 for RPE. Bland-Altman plots showed a bias of 0.50% for HRreserve, 3.2% for VO2max, 0.05 for affect, and -0.35 for RPE. Conclusion: Self-selected intensity during aerobic exercise performed on treadmill is reliable, which promotes adequate and lower MDC values on physiological and perceptual responses in elderly women.