High levels of genetic diversity and connectivity are crucial for the persistence of local populations, especially at the edge of species' distribution ranges. Here, we assessed the potential and realized connectivity of populations of the Mediterranean seagrass Posidonia oceanica at its easternmost distribution using physical modelling and genetic analyses. Genetic assessments of diversity and gene flow among populations were carried out with 18 microsatellite loci, while oceanographic connectivity was assessed via Lagrangian dispersal simulations. Levels of genetic and clonal diversities were prevalent among shallow and deep sites without signs of reproductive isolation. Both approaches identified two main clusters corresponding to "Aegean" populations along the western Turkey coast and "Levantine" populations along the southern Turkey coast. Aegean populations were genetically homogeneous, connected by high levels of gene flow, whereas Levantine populations were genetically heterogeneous. Within-sea patterns of genetic connectivity did not fully overlap with those derived from physical modelling; the realized connectivity was greater than that predicted by ocean-current simulations, especially in the Aegean Sea. Lagrangian dispersion dynamics cannot necessarily explain genetic connectivity patterns among populations, which are shaped over longer temporal scales and can be affected by human activities and local environmental conditions.