Tectonic reconstructions of Southeast Asia have given rise to numerous controversies that include the accretionary history of Sundaland and the enigmatic tectonic origin of the proto-South China Sea. We assimilate a diversity of geological and geophysical observations into a new regional plate model, coupled to a global model, to address these debates. Our approach takes into account terrane suturing and accretion histories, the location of subducted slabs imaged in mantle tomography in order to constrain the evolution of regional subduction zones, as well as plausible absolute and relative plate velocities and tectonic driving mechanisms. We propose a scenario of rifting from northern Gondwana in the latest Jurassic, driven by northward slab pull from north-dipping subduction of Tethyan crust beneath Eurasia, to detach East Java, Mangkalihat, southeast Borneo and West Sulawesi blocks that collided with a Tethyan intraoceanic subduction zone in the mid-Cretaceous and subsequently accreted to the Sunda margin (i.e., southwest Borneo core) in the Late Cretaceous. In accounting for the evolution of plate boundaries, we propose that the Philippine Sea plate originated on the periphery of Tethyan crust forming this northward conveyor. We implement a revised model for the Tethyan intra-oceanic subduction zones to reconcile convergence rates, changes in volcanism and the obduction of ophiolites. In our model the northward margin of Greater India collides with the Kohistan-Ladakh intra-oceanic arc at ∼ 1000 km beneath northern Borneo and the South China Sea are likely to be remnants of the proto-South China Sea basin.