Personalized medicine holds tremendous promise for improving safety and efficacy of drug therapies by optimizing treatment regimens. Rapidly developed patient-derived xenografts (pdx) could be a helpful tool for analyzing the effect of drugs against an individual's tumor by growing the tumor in an immunodeficient animal. Severe combined immunodeficiency (SCID) mice enable efficient in vivo expansion of vital tumor cells and generation of personalized xenografts. However, they are not amenable to large-scale rapid screening, which is critical in identifying new compounds from large compound libraries. The development of a zebrafish model suitable for pdx could facilitate large-scale screening of drugs targeted against specific malignancies. Here, we describe a novel strategy for establishing a zebrafish model for drug testing in leukemia xenografts. We used chronic myelogenous leukemia and acute myeloid leukemia for xenotransplantation into SCID zebrafish to evaluate drug screening protocols. We showed the in vivo efficacy of the ABL inhibitor imatinib, MEK inhibitor U0126, cytarabine, azacitidine and arsenic trioxide. We performed corresponding in vitro studies, demonstrating that combination of MEK- and FLT3-inhibitors exhibit an enhanced effect in vitro. We further evaluated the feasibility of zebrafish for transplantation of primary human hematopoietic cells that can survive at 15 day-post-fertilization. Our results provide critical insights to guide development of high-throughput platforms for evaluating leukemia.