This study compared the parameter estimates of pulmonary oxygen uptake (VO(2p)), heart rate (HR) and muscle deoxygenation (Δ[HHb]) kinetics when several moderate-intensity exercise transitions (MODs) were performed during a single visit versus several MODs performed during separate visits. Nine subjects (24 ± 5 years, mean ± SD) each completed two successive cycling MODs on six occasions (1-6A and 1-6B) from 20 W to a work rate corresponding to 80% estimated lactate threshold with 6 min recovery at 20 W. During one visit, subjects completed two series of three MODs (6A-F), separated by 20 min rest. VO(2p) time constants (τVO(2p); 27 ± 10 s, 25 ± 12 s, 25 ± 11 s) were similar (p > 0.05) for MODs 1-6A, 1-6B and 6A-F, respectively. τVO(2p) had reproducibility 95% confidence intervals (CI(95)) of 8.3, 8.2, 4.7, 4.9 and 4.7 s when comparing single (1A vs. 2A), the average of two (1-2A vs. 3-4A), three (1-3A vs. 4-6A), four (1-2AB vs. 3-4AB) and six (1-3AB vs. 4-6AB) MODs, respectively. The effective Δ[HHb] response time (τ'Δ[HHb]) was unaffected across conditions (1-6A: 19 ± 2 s, 1-6B: 19 ± 3 s, 6A-F: 17 ± 4 s) with reproducibility CI(95) of 5.3, 4.5, 3.1, 2.9 and 3.3 s when a single, two, three, four and six MODs were compared, respectively. τHR was reduced in MODs 6A-F compared to 1-6A and 1-6B (23 ± 5 s, 25 ± 5 s, 27 ± 6 s, respectively). This study showed that parameter estimates of VO(2p), HR and Δ[HHb] kinetics are largely unaffected by data collection sequence, and the day-to-day reproducibility of τVO(2p) and τ'Δ[HHb] estimates, as determined by the CI(95), was appreciably improved by averaging of at least three MODs.