Osteochondral defects (i.e., defects which affect both the articular cartilage and underlying subchondral bone) are often associated with mechanical instability of the joint, and therefore with the risk of inducing osteoarthritic degenerative changes. Current surgical limits in the treatment of complex joint lesions could be overcome by grafting osteochondral composite tissues, engineered by combining the patient's own cells with three-dimensional (3D) porous biomaterials of pre-defined size and shape. Various strategies have been reported for the engineering of osteochondral composites, which result from the use of one or more cell types cultured into single-component or composite scaffolds in a broad spectrum of compositions and biomechanical properties. The variety of concepts and models proposed by different groups for the generation of osteochondral grafts reflects that understanding of the requirements to restore a normal joint function is still poor. In order to introduce the use of engineered osteochondral composites in the routine clinical practice, it will be necessary to comprehensively address a number of critical issues, including those related to the size and shape of the graft to be generated, the cell type(s) and properties of the scaffold(s) to be used, the potential physical conditioning to be applied, the degree of functionality required, and the strategy for a cost-effective manufacturing. The progress made in material science, cell biology, mechanobiology and bioreactor technology will be key to support advances in this challenging field.