The state of stress in the lithosphere provides strong constraints on the forces acting on the plates. The directions of principal stresses in the plates as indicated by midplate earthquake mechanisms, in situ stress measurements, and stress‐sensitive geological features are used to test plate tectonic driving force models, under the premises that enough data exist in selected areas to define a regionally consistent stress field and that for most such areas the dominant forces producing such stresses are plate tectonic in origin. Force models include buoyancy forces at ridges, subduction zones, and continental convergence zones and variously parameterized viscous shear between the lithosphere and the asthenosphere. A linear finite element method, based on the wave front solution technique, is used to predict the intraplate stress for each force model. Several long‐wavelength patterns for the orientation of horizontal principal deviatoric stresses are observable in the stress data. Maximum compressive stresses trend E‐W to NE‐SW for much of stable North America and E‐W to NW‐SE for continental South America. In western Europe the maximum compressive stresses trend NW‐SE, while in Asia the trend is more nearly N‐S, especially near the Himalayan front. In the Indian plate the trend varies from nearly N‐S in continental India to more nearly E‐W in Australia. Horizontal stresses are variable in Africa but tend to indicate a NW‐SE trend for the maximum compressive stress in west Africa and an E‐W trend for the minimum compressive stress in east Africa. Oceanic lithosphere away from plate boundaries is generally in a state of deviatoric compression, although few focal mechanisms can be constrained to define the orientation of the principal stresses.Comparison of stress orientations predicted for a wide range of driving force models to these regional stress observations provides a powerful test of the models. Ridge pushing forces are required in all models that match the stress orientation field. The net pulling force of subducted lithosphere, if such a force acts approximately symmetrically about the plate boundary, is at most a few times larger than other forces acting on the plates. Resistive forces associated with trench thrust faults and motion of the slab with respect to the mantle must therefore nearly balance the large gravitational potential of the slab. The upper limit on the ratio of net slab to ridge forces may be increased by less than a factor of 2 if net slab forces are reduced for the fastest moving plates by assuming that the resistance to subduction increases with convergence rate. Forces acting to resist further continental convergence along the Alpine‐Himalayan belt are important for models of the intraplate stress field in Europe, Asia, and the Indian plate. Inclusion of continental convergence zone forces, however, does not affect the upper bound on the ratio of net slab to ridge forces. A variety of possible lithosphere‐asthenosphere interactions have been tested. Resistive viscous drag forces acting on the base of the lithosphere improve the fit between calculated and observed stresses in the Nazca and South American plates as long as the drag coefficient is nonzero beneath oceanic lithosphere. The calculated intraplate stress field is not very sensitive to an increased drag coefficient beneath old oceanic lithosphere compared to young oceanic or continental lithosphere. Increasing the drag coefficient beneath continents compared to oceanic lithosphere by a factor of 5 or 10 has little effect on the overall fit of calculated stresses to observed stresses. Models in which viscous drag forces drive, rather than resist, plate motions are in poor agreement with intraplate stress data, although this lack of fit may depend on the oversimplified model of mantle flow patterns that has been assumed. A model of the driving mechanism in which slab forces act only on the plate with subducted lithosphere and where viscous drag forces are assumed to balance the torque on each plate produced by ridge and trench forces predicts stresses in good agreement with the data for most continental regions. The fit between calculated and observed stresses for this model is relatively poor for most oceanic regions. This model suggests, however, that it is probably an oversimplification to assume that the net force exerted by the slab acts symmetrically about the plate boundary, though some pulling force on the overthrust plate appears necessary. While the role of drag forces in the driving mechanism remains poorly constrained, the finite element technique that has been developed may be applied in the future to any specific and realizable flow pattern that is proposed for the mantle.