In this work we use in silico tools like de novo drug design, molecular docking and absorption, distribution, metabolism and excretion (ADME) studies in order to develop new inhibitors for tyrosine-kinase protein (including its mutate forms) involved in myeloid leukemia disease. This disease is the first cancer directly associated with a genetic abnormality and is associated with hematopoietic stem cells that are manifested primarily with expansion myelopoiesis. Starting from a family of fragment and seeds from known reference drugs, a set of more than 6k molecules were generated. This first set was filtered using the Tanimoto similarity coefficient as criterion. The second set of more dissimilar molecules were then used in the docking and ADME studies. As a result, we obtain a group of molecule that inhibit the tyrosine-kinase family and have ADME properties better than the reference drugs used in the treatment of myeloid leukemia.