The structure of the eastern Sunda back arc region is dominated by two large north directed thrusts, the Wetar and Flores thrusts, and one or more minor thrusts, which may represent early stages of subduction polarity reversal of the arc. The relationships between thrusting and magmatic activity, surface slopes, cross-arc faulting, uplift, forearc structure, and collision by the Australian continental margin provide constraints of the likely driving mechanisms of thrust formation. Driving mechanisms are considered in two broad categories. One set of possible mechanisms originates within the upper plate and includes gravity sliding, gravity spreading, and magmatic intrusion. The other set focuses on stress propagation between the upper and lower plates, and for this arc system the major mechanism is that of continental collision. While no one mechanism explains the range of observations we see in the back arc region, all of these aspects play some role in the development of the thrusts. We can eliminate upper plate mechanisms as primary driving forces because slope effects are similar in areas of both thrusting and nonthrusting. In addition, the Wetar thrust shows a negative relation between thrusting and volcanism, and side-looking airborn radar images show no evidence of postvolcanic rifting that could accompany intrusion without volcanism. Collision between the arc and the Australian continent provides the clearest driving mechanism, but it does not explain the discontinuous nature or the particular location of the thrusts. We suggest a sequence that appears to account for the data and to provide insight into the controversies often associated with other such thrust belts. Thrusting was initiated in those areas where the crust of the forearc region was thick, a factor that facilitated stress propagation across the arc, and where the crust of the back arc was thin, which facilitated back arc thrusting. Thrusting concentrated at the base of the slope where the slope stress is maximum. The early thrusting may have been helped also by thermal weakening of the crust due to volcanism. As convergence proceeded, volcanism waned then ceased in the eastern part of the arc. The crust then strengthened, and thrusting required higher stress to maintain and propagate. This later stage should show higher seismic energy release, which appears to be the case for the Wetar thrust region. Displacement on the Flores thrust appears to be matched by deflection of the volcanic arc. We have estimated shortening in the back arc from interpretation of reflection profiles, and for the Flores thrust the value of approximately 30 km is consistent with that estimated for deflection of the arc. Compressional deformation continues west of the collision zone, across the Bali Basin and into Java. It is reported in NE Sumatra as well. This deformation involves much less convergence than the eastern Sunda thrusts, and it may be related to magmatic processes, to subduction of oceanic plateaus, such as the Roo Rise, or, in the case of the Bali Basin, to efficient lateral propagation of the Flores thrust.