A paleomagnetic and geochronologic study was carried out on Late Permian-to-Early Triassic magmatic units exposed in the North Patagonian Massif, near the locality of Estancia La Esperanza (Rio Negro Province, Argentina), to provide better paleogeographic and tectonic constraints on the evolution of Patagonia and its relations with Gondwana in the Late Paleozoic–Early Mesozoic. The study included the Late Permian (264 ± 2 Ma) Rhyolite Dome, the Collinao Dacite, for which we obtained a U–Pb zircon crystallization age of 253 ± 2 Ma, several basic dikes, that intrude older granite units (255 ± 2 Ma), and a swarm of acidic dikes for which an Early Triassic age had been previously assigned that we confirmed with a U–Pb zircon age of 244 ± 2 Ma. A paleomagnetic pole (C), computed on the basis of seven VGPs, was obtained for the Collinao Dacite and associated units (48.3°S, 349.9°E, N = 7, A95 = 15.1°). A single VGP for the somewhat older Rhyolite Dome falls close to that pole. A second paleomagnetic pole (A) which strongly disagrees with the C position was computed based on 12 VGPs from the same number of acidic dikes (87°S, 51.5°E, N = 12, A95 = 8.1°). The basic dikes, with ages bracketed between the dacitic and the rhyolitic dikes, yielded four VGPs, two of which are consistent with the C pole and the other two are close to the A pole. The C pole falls on a much older position (Late Carboniferous) on different Gondwana/South America reference paths, which is clearly anomalous based on its Late Permian age. The A pole, in turn, is consistent with the 250/240 Ma paleomagnetic poles in most reference paths for Gondwana/South America. Either regional tilting or rotation around a vertical axis to explain the anomalous C pole is unlikely considering the local geological evidence. A working model is presented to explain the discrepancy as due to around 30° CCW rotation of the whole North Patagonian Massif during the latest Permian and earliest Triassic, closing a V-shaped basin that separated it from southern Gondwana, as the final stages of collision of an allochthonous or para-autochthonous Patagonia terrane. This hypothesis is consistent with Late Permian-to-Early Triassic evidence of deformation in the Hesperides Basin as well as in northern Patagonia. Scarce previous paleomagnetic data from this region can be reconciled with this model, considering that they are very loosely constrained in age.