Learning performance is influenced by emotional reactivity, low reactivity being generally beneficial. Previous experiments show that emotional reactivity can be modified after a period of social isolation. We hypothesized that eleven days of isolation would affect yearlings' emotional reactivity and improve their learning abilities. Twenty-five yearlings were divided into two groups: 12 were continuously isolated for 11 days (isolated) and 13 stayed together (control). During the period of isolation, all yearlings underwent two learning tasks: a habituation procedure in which a novel object was presented for 120s every day, either when the horse was alone (isolated) or with conspecifics (control); an instrumental learning task in which the yearling had to walk forwards or backwards to obtain a food reward. At the end of the isolation period, animals performed tests to assess aspects of emotional reactivity: reactivity to novelty, to humans, to social separation, to suddenness and to sensory stimuli. Results showed that isolated yearlings habituated more to the novel object than controls and performed better in the instrumental task. Moreover, they were less reactive to novelty, to social separation and to suddenness than controls. Overall, these data suggest that the better performance of isolated yearlings could be explained by a decrease in their emotional reactivity.