Autologous bone marrow transplantation (ABMT) was introduced as a treatment for terminal leukemic relapse more than 40 years ago. For childhood acute lymphoblastic leukemia (ALL) the role of ABMT is well defined. Some studies suggest that it is also beneficial for adult patients with high-risk factors or with relapse. However, these inferences are based on a relatively small number of patients with short follow-up. Nevertheless patients with high-risk ALL are candidates for ABMT if no histocompatible sibling is available. Similarly patients in second or later complete remission (CR) in the absence of a histocompatible donor may derive benefit from ABMT. The different conditioning regimens used by the treatment centers are associated with different toxicities but none has been proven to be superior than others. In the majority of studies the marrow has been purged of leukemic cells, but this maneuver has never been evaluated in a randomized comparative trial. Transplant related mortality rate of ABMT is low compared to allogeneic transplantation. The GvL effect, which is important to eliminate malignant cells in acute and chronic myelocytic leukemia, has not been definitively demonstrated in ALL. The tyrosine-kinase inhibitor STI 571 offers new perspectives for patients with the Phl/bcr/abl translocation. It may be especially useful for treating minimal residual disease (MRD) before and/or after ABMT.