Two divergent behavioural and physiological response patterns to challenges have been identified in mammals and birds, frequently termed the proactive and reactive coping styles. In recent years, individually distinct coping styles have also been observed in several species of fish. These individual differences may result in suboptimal production and compromised welfare in aquaculture. An approach to overcome these problems could be to sort fish and optimise rearing conditions according to coping style. It has been previously demonstrated that the proactive and reactive coping styles in fish can be characterised by contrasting behavioural responses to hypoxia. Two rainbow trout (Oncorhynchus mykiss) strains, bred for a low- (LR) and high- (HR) cortisol response to a standardized stressor, are suggested to resemble the proactive and reactive coping styles respectively. Therefore, these fish provided an opportunity for verifying a method for sorting fish with respect to coping style by exposure to hypoxia. Groups consisting of 24 individually tagged fish, 12 HR and 12 LR were exposed to hypoxia in a two choice system. The system consisted of a “home” tank provided with cover connected to a second brightly “illuminated” tank via a closable doorway. During the experiment, the doorway between the two tanks was opened and hypoxic conditions were gradually induced in the “home” tank by bubbling with nitrogen. The latency time to move away from hypoxic conditions to normoxic conditions in the second tank was recorded for each individual. The oxygen saturation in each tank was measured every 30 min. The experiment consisted of two trials. Each trial was carried out in two sessions, switching the “home” tank and “illuminated” tank between Sessions 1 and 2. The results indicated that the response to hypoxia differed significantly between LR and HR individuals in both Session 1 (P <0.05) and Session 2 (P ≤ 0.001). Furthermore, a higher number of HR individuals left hypoxic conditions compared to LR individuals in both Session 1 (P = 0.001) and Session 2 (P ≤ 0.001). Taken together, the findings of the present study demonstrate a repeatable difference in behavioural response to hypoxia between the two strains. The method presented could be utilized as a non-invasive method for sorting fish according to stress coping style.