Paleoseismological trenching studies constrain recurrence times and magnitudes of faulting events and earthquakes on active faults. In a trench along the central part of the Peel Boundary fault zone (PBFZ), southeastern Netherlands, evidence was found for such a large faulting event that occurred around 14 ka. The event caused a fault scarp in unconsolidated sediments of ∼1 m height. A colluvial wedge was formed next to the scarp. A second faulting event offsets this colluvial wedge by 0.2–0.1 m. This event can be tentatively dated at ∼13 ka. During or immediately after the second event, a large clastic dyke intruded along the fault plane. The dyke is not faulted, but its emplacement did cause some minor thrust faulting around the injection. The sudden character of the main faulting event, the brittle deformation style of loam layers, the lack of growth faulting in the colluvial wedge, the clastic dykes and the flame structures demonstrate that the main faulting event was a surface rupturing earthquake. Based on the scarp height, the estimated moment magnitude is about 6.8 ± 0.3. Similar observations in a previous trench site suggest that the length of the surface rupture was at least 32 km. The earthquake took place during the Weichselian (Würmian) Late Glacial. This timing corresponds to the start of the glacio-isostatic forebulge collapse in the Netherlands. Glacio-isostatic movements have been invoked before to explain earthquake events in the Roer Valley Rift System in which the PBFZ is situated, and in northern Germany and Denmark. If these earthquakes can indeed be attributed to a collapsing forebulge, their ages should show a decrease in the direction of ice-sheet retreat. This might indeed be the case, as the ages decrease from 14 ka and 13 ka in this trench via 13–16 ka at the Osning Thrust Zone, NW Germany, to 12–14 ka in northern Denmark.