Purpose: To characterize the relationships between muscle oxygenation and performance during on- and off-water tests in highly trained sprint canoe–kayak athletes. Methods: A total of 30 athletes (19 kayakers and 11 canoeists) performed a maximal incremental test on a canoe or kayak ergometer for determination of VO2max and examination of the relation between peak power output (PPO) and physiological parameters. A subset of 21 athletes also performed a 200- and a 500- (for women) or 1000-m (for men) on-water time trial (TT). Near-infrared spectroscopy monitors were placed on the latissimus dorsi, biceps brachii, and vastus lateralis during all tests to assess changes in muscle O2 saturation (SmO2) and deoxyhemoglobin concentration ([HHb]). The minimum O2 oxygenation (SmO2min) and maximal O2 ($Δ$[HHb] extraction) were calculated for all subjects. Results: PPO was most strongly correlated with VO2max (R = .9), but there was also a large correlation between PPO and both SmO2min and $Δ$[HHb] in latissimus dorsi (R = −.5, R = .6) and vastus lateralis (R = −.6, R = .6, all P < .05). Multiple regression showed that 90% of the variance in 200-m performance was explained by both $Δ$[HHb] and SmO2min in the 3 muscles combined (P < .01) and 71% of the variance in 500-/1000-m performance was explained by $Δ$[HHb] in the 3 muscles (P < .01). This suggests that O2 extraction is a better predictor of performance than VO2max in sprint canoe–kayak. Conclusions: These results highlight the importance of peripheral adaptations in both short and long events and stress the relevance of adding muscle oxygenation measurements during testing and racing in sprint canoe–kayak.