Many organisms perform regular migrations over long distances. These movements are often related to feeding and reproductive periods and regulated by oceanographic conditions as well as physiological and behavioural traits. Different individual traits and their associated evolutionary constraints will ultimately shape the migratory strategy (and route) of individuals. Optimality theory can provide a framework to assess these inherent trade-offs in individual migrations and identify optimal migration routes in different conditions. Here, we present a model that describes the behavioural trade-off between migration time and energy expenditure and identifies optimal migration routes in realistic ocean conditions. The model explicitly includes a behavioural factor for individual risk management, including risks associated with moving in a stochastic environment. We test this model in three different case studies, one in an idealized theoretical context and two in realistic conditions for sea turtle migrations. We show that behavioural traits can largely influence the optimal routes in long-distance migrations, resulting in major changes in migratory pathways. Further, we assess the ability of the model to infer back behavioural traits given a set of synthetic individual tracks and show relatively good performances. However, further tests are needed to evaluate performances when accurate observations of migrations are used.