Effective management of exploited populations is based on an understanding of population dynamics and evolutionary processes. In spatially structured populations, dispersal is a central process that ultimately can affect population growth and viability. It can be influenced by environmental conditions, individual phenotypes, and stochastic factors. However, we have a limited knowledge of the relative contribution of these components and its interactions, and which traits can be used as reliable predictors of the dispersal ability. Here, we conducted a longitudinal field experiment aimed to identify traits which can be used as proxy for dispersal in juvenile brown trout (Salmo trutta L.). We measured body size and standard metabolic rates, and estimated body shapes for 212 hatchery-reared juvenile fish that were marked with individual codes and released in a small coastal stream in northwest Spain. We registered fish positions and distances to the releasing point after 19, 41, 60 and 158 days in the stream. We detected a high autocorrelation of dispersal distances, demonstrating that most individuals settle down relatively soon and then hold stable positions over the study period. Body size and fish shape were reliable predictors of dispersal, with bigger and more robust-set individuals being more likely to settle closer to the release site than smaller and more elongated fish. In addition, the analysis of spacing and spatial patterns indicated that the dispersal of introduced fish could affect the distribution of resident conspecifics. All together, these results suggest that stocking programs aimed to the enhancement of overexploited populations at fine spatial scales can be optimized by adjusting the size and shape of the introduced fish to specific management targets and environmental conditions.