Numerous studies have been conducted in an attempt to discover cooling strategies that can be effective in improving exercise performance. However, the mechanism by which neck cooling relieves exercise-induced physiological stress and the optimal cooling temperature are unclear. This study aimed to investigate the effects of neck cooling at different temperatures during high-intensity cycling on body temperature, physiological variables, oxidant/antioxidant status, heat shock protein (HSP) 70 levels, and exercise performance in adolescent athletes. Seven well-trained male adolescent cyclists (age, 17.00 ± 0.76 years; athletic career, 3.86 ± 0.90 years) participated in three exercise trials involving three cooling regimens: control (CON), low-temperature (7 °C) neck cooling (LNC), and mixed-temperature (14 + 20 °C) neck cooling (MNC). The experimental condition used a cross-over design to minimize adaption to the repetitive cycling trials. Cycling consisted of a 20 km warm-up session and a two 2 km race session. Neck cooling at different temperatures was administered for 20 min during each rest period: after the warm-up, after the first 2 km race, and after the second 2 km race. Blood samples were taken to assess serum malondialdehyde (MDA), superoxide dismutase (SOD), and HSP70 levels. In addition, tympanic temperature (Tty), thermal sensation (TS), heart rate (HR), and the saturation of percutaneous oxygen (SpO2) were measured before, immediately after, and 24 h after exercise. As a measure of cycling performance, the race record and speed were measured in the first and second 2 km races. In all trials, Tty, TS, HR, MDA, SOD, and HSP70 levels significantly increased (p < 0.05), and SpO2 levels significantly decreased (p < 0.05). TS significantly decreased 24 h after exercise compared to that immediately after exercise in the LNC and MNC trials (p < 0.05). Serum HSP70 levels were significantly higher 24 h after exercise (0.87 ± 0.10 ng/mL) than immediately after exercise (0.79 ± 0.04 ng/mL) in the CON trial (p < 0.05). Twenty-four hours after exercise, the CON (0.87 ± 0.10 ng/mL) trial showed significantly higher serum HSP70 levels than the LNC (0.73 ± 0.01 ng/mL) trial (p < 0.05). There was no significant difference in cycling race record or speed between the trials (p > 0.05). Our findings suggest that neck cooling can induce a positive effect on thermal perception during recovery after cycling and that neck cooling at a relatively low temperature may be more effective in reducing exercise-induced HSP70 expression.