The Mexican volcanic belt has been traditionally regarded as a classic case of subduction-related calc-alkalic volcanism. However, a series of geologic, geophysical, and petrological arguments makes this simple relationship doubtful. A seismic gap beneath the belt, a large-scale mantle anomaly, a graben triple-junction domain, and the presence of volumetrically important oceanic-island basalt (OIB) volcanism throughout the belt suggest a more complex tectonic scenario involving plume- and subduction-related processes. We here propose a model involving the development of a propagating rift opening from west to east in response to plume activity. The process started in Miocene time within the western sector of the belt (Guadalajara) and gave rise to a graben triple junction and OIB-type and calc-alkalic volcanism. Extension and volcanism proceeded to the east, giving rise to progressively younger ages for the initiation of OIB-type volcanism: (1) Miocene in the west (e.g., Guadalajara), (2) Pliocene in the central zone (e.g., Michoacán-Guanajuato), and (3) Quaternary farther east (e.g., Chichinautzin). Geochemical evidence suggests that part of the modern calc-alkalic volcanism (e.g., Chichinautzin) may be derived from magma mixing between the OIB mafic magmas and silicic, crust-derived magmas. However, we do not preclude some influence of the subducting slab in the generation of other (e.g., Jorullo) calc-alkalic volcanic rocks. Our model suggests a currently unrooted upper plume attached to the subcontinental lithosphere, which defines a hot zone beneath the Mexican volcanic belt.