The duration of breath-hold dives and the available time for foraging in submerged prey patches is ultimately constrained by oxygen balance. There is a close relationship between swim speed and oxygen utilisation, so it is likely that breath-holding divers optimise their speeds to and from the feeding patch to maximise time spent feeding at depth. Optimal foraging models suggest that transit swim speed should decrease to minimum cost of transport (MCT) speed in deeper and longer duration dives. Observations also suggest that descent and ascent swimming mode and speed may vary in response to changes in buoyancy. We measured the swimming behaviour during simulated foraging of seven captive female grey seals (two adults and five pups). Seals had to swim horizontally underwater from a breathing box to a submerged automatic feeder. The distance to the feeder and the rate of prey food delivery could be varied to simulate different feeding conditions. Diving durations and distances travelled in dives recorded during these experiments were similar to those recorded in the wild. Mean swim speed decreased significantly with increasing distance to the patch, indicating that seals adjusted their speed in response to travel distance, consistent with optimality model predictions. There was, however, no significant relationship between the transit swim speeds and prey density at the patch. Interestingly, all seals swam 10-20% faster on their way to the prey patch compared to the return to the breathing box, despite the fact that any effect of buoyancy on swimming speed should be the same in both directions. These results suggest that the swimming behaviour exhibited by foraging grey seals might be a combination of having to overcome the forces of buoyancy during vertical swimming and also of behavioural choices made by the seals.