We addressed the importance of defining a mechanical testing methodology for the compression of human trabecular bone specimens. In fact, currently there are several protocols to test trabecular bone, but a single, standard and validate method has not been accepted yet. In our work, human femoral epiphyses collected from patients with osteoporosis (fragility fractures) and hip osteoarthritis, submitted to total hip replacement surgery, were used. The aims of our work were to develop a mechanical testing methodology for the compression of trabecular bone specimens taking into account the optimization of bone extrinsic and intrinsic variables, in order to establish a patient bone sample database with clinical, structural and mechanical information. Extrinsic variables, such as the effect of specimen preparation, with particular focus on the dimensions of test specimens, and others associated with the compression test, such as the method employed to determine specimen deformation, and hence strain, were evaluated. Also, a new device used to withhold the specimens was developed and tested by the present authors. Although each specimen showed a unique behaviour, even when comparing compression curves between patients from the same disease group, implicating additional complexity and difficulty in the data analysis, the authors managed to assemble the results in two groups related with a possible difference in the deformation mechanisms occurring after yielding.