This study compared the acute effects of three recovery methods: active recovery (AR), hot- and cold-water immersion (HWI and CWI, respectively), used between two training sessions in elite athletes. Twelve national-team skaters (7 males, 5 females) completed three trials according to a randomized cross-over study. Fifteen minutes after an exhaustive ice-skating training session, participants underwent 20 min of HWI (41.1 ± 0.5°C), 15 min of CWI (12.1 ± 0.7°C) or 15 min of active recovery (AR). After 1 h 30 min of the first exercise, they performed a repeated-sprint cycling session. Average power output was slightly but significantly higher for AR (767 ± 179 W) and HWI (766 ± 170 W) compared to CWI (738 ± 156 W) (p = 0.026, d = 0.18). No statistical difference was observed between the conditions for both lactatemia and rating of perceived exertion. Furthermore, no significant effect of recovery was observed on the fatigue index calculated from the repeated sprint cycling exercises (p > 0.05). Finally, a positive correlation was found between the average muscle temperature measured during the recoveries and the maximal power output obtained during cycling exercises. In conclusion, the use of CWI in between high-intensity training sessions could slightly impair the performance outcomes compared to AR and HWI. However, studies with larger samples are needed to confirm these results, especially in less trained athletes.