The use of animal models is a very powerful tool for the preclinical assessment of potential therapies for osteoporosis. However, the effective use of animal models has two prerequisites. The first is the use of appropriate techniques to assess the overall effects of therapy on bone. As spontaneous fractures do not occur in any species other than humans, the efficacy of a therapy cannot be assessed by its impact on fracture incidence. Instead, a suite of parameters (collectively referred to as 'bone quality'), including bone architecture, mineralization and mechanical properties, is examined. While techniques such as histomorphometry and dual-energy x-ray absorptiometry are useful for elucidating specific effects of therapies on bone, mechanical testing is crucial to integrate all the changes into parameters which can serve as a proxy for fracture risk. The second prerequisite is an understanding of the similarities and differences between various animal models of osteoporosis and the human case. Two important animal models of postmenopausal osteoporosis are discussed here: the ovariectomized (OVX) rat and the non-human primate (NHP) model. The OVX rat is convenient, inexpensive and responds rapidly to ovariectomy and to therapy at a number of cancellous sites, which is congruent to the human case. However, the OVX-induced changes are largely confined to cancellous bone. Non-human primates, however, are biologically much more similar, particularly in terms of their reproductive physiology, to humans. The response of cortical bone of skeletally mature NHPs to ovariectomy and other stimuli closely parallels that in humans, but this is not mirrored in the cancellous bone in younger animals. This suggests that these two animal models complement each other as models for osteoporosis in humans.