Seagrasses worldwide are noted for suffering from mechanical damage caused by boat anchoring. This is particularly so in sites highly frequented by boaters (marine protected areas or coastal urbanised areas). In the last decades, different strategies have been put into practice to reduce such impacts on seagrasses (i.e. by anchoring bans or by deploying boat moorings). More recently, in consideration that few marine protected area (MPA) management bodies or local administrations have the resources to enforce their anchorage regulations, the self-regulatory approach based on education and information of boaters has been preferred in several cases. At present, however, very little is known on the correct anchoring practices to ensure the safeguarding of seagrasses. The aim of the present study was to experimentally quantify in the field the damage caused to Posidonia oceanica shoot density by anchoring. A multifactorial experiment was designed to test whether the damage is dependent on (1) different anchor types (Hall, Danforth and Folding grapnel), (2) the use of a chain vs. a rope, (3) the three anchoring stages (anchor fall, dragging/lock-in and weighing), and finally (4) whether the pattern is consistent among different locations of the meadow.As expected, the three anchor types employed in the present study differed in the levels of damage inflicted on the P. oceanica meadows of the Ustica Island MPA. In particular, the use of the Hall type anchor seems to be preferable to minimise this impact in comparison with the other two anchor types. Moreover, the effect on the meadow of the three anchor types is greatly dependent on the anchoring stage. These results confirm that the weighing stage is the critical stage of the anchoring process. The number of damaged shoots of P. oceanica was not affected by the presence of the chain. These patterns were consistent between locations.