Dispersal ability is a key determinant of the propensity of an organism to cope with habitat fragmentation and climate change. Here we quantify queen dispersal in two common bumblebee species in an arable landscape. Dispersal was measured by taking DNA samples from workers in the spring and summer, and from queens in the following spring, at 14 sites across a landscape. The queens captured in the spring must be full sisters of workers that were foraging in the previous year. A range of sibship reconstruction methods were compared using simulated data sets including or no genotyping errors. The program Colony gave the most accurate reconstruction and was used for our analysis of queen dispersal. Comparison of queen dispersion with worker foraging distances was used to take into account an expected low level of false identification of sister pairs which might otherwise lead to overestimates of dispersal. Our data show that Bombus pascuorum and B. lapidarius queens can disperse by at least 3 and 5 km, respectively. These estimates are consistent with inferences drawn from studies of population structuring in common and rare bumblebee species, and suggest that regular gene flow over several kilometres due to queen dispersal are likely to be sufficient to maintain genetic cohesion of ubiquitous species over large spatial scales whereas rare bumblebee species appear unable to regularly disperse over distances greater than 10 km. Our results have clear implications for conservation strategies for this important pollinator group, particularly when attempting to conserve fragmented populations.