With the aim to better understand the timing and geological evolution of the late Cenozoic stratigraphic record of the onshore, western Valdés Basin in northeast Chubut province (Patagonia, Argentina), we perform sedimentological, stratigraphic, and zircon U–Pb geochronologic analyses. The study includes the terrestrial Early Miocene Trelew Member of the Sarmiento Formation, the shallow marine Early Miocene Gaiman Formation, the shallow marine to estuarine Late Miocene Puerto Madryn Formation, and the Plesitocene fluvial conglomerates of the Rodados Patagónicos. Based on our data set, together with the recognition of relevant stratigraphic surfaces, we divide the succession into stratigraphically bounded units and define three cycles of sedimentation. The Lower Cycle rests on top of Surface 1, which comprise a long hiatus including the entire Oligocene Epoch. This cycle is in turn divided into two subunits separated by the transgressive, ravinement Surface 2. The Lower Cycle is separated from the overlying Middle Cycle (=Puerto Madryn Formation) by the flat and regionally extensive Surface 3. Finally, the latter is separated from the Upper Cycle (=Rodados Patagónicos) by Surface 4, a complex surface at the base of gravel deposits forming several terrace levels. Maximum depositional ages calculated from 10 zircon U–Pb samples indicate the succession between surface 1 and 4 extends from 21.5 to 2.7 Ma (Aquitanian-Piacenzian). The Lower Cycle is restricted to the 21.5–16 Ma interval (Aquitanian-Burdigalian); the Middle Cycle between 12 and 2.7 Ma (latest Serravalian-Piacenzian), with most of the new and previous ages clustering in the Tortonian; whereas the Upper Cycle is restricted to the Pleistocene. Although most of the analyzed sediments show fresh volcanic glass particles, older zircon ages are dominant, representing detrital components with a dominant peak of Permian-aged grains, with subordinated Devonian, Jurassic, Cretaceous, and Paleogene peaks. We interpret the analyzed succession as a condensed section compared to the offshore Valdés Basin depocenter area. Sedimentation in the study area was controlled by periods of slow subsidence, probably related to thermal subsidence and sediment compaction in the depocenter. Two main sources of sediments for the analyzed succession are envisaged: one represented by explosive volcanism located either in the Andes or in the North Patagonian Massif, the other through erosion of older rocks surrounding the North Patagonian Massif where late Paleozoic plutonic rocks are abundant. Particularly for the Middle and Upper cycles, an increase in Mesozoic contribution suggests erosion from sediments of the Cañadón Asfalto Basin. The new ages and stratigraphic analyses here reported are of major relevance for the integration of the onshore and offshore stratigraphies of the Valdés Basin which is key to connect the geologic evolution of the Andean and passive margin areas of southern South America. They are also of major relevance for the calibration of fossil communities, particularly for the abundant and diverse marine vertebrates preserved in the Burdigalian Gaiman Formation.