The importance of perturbation in transforming growth factor beta (TGFβ) signaling for the onset and progression of cancer is well established. Many tumors over express TGFβ, and high circulating levels of TGFβ1 in cancer patients are frequently associated with poor prognosis. TGFβ has context-dependent biphasic action during tumorigenesis. Because of this, it is essential to take due care about the selection of patients most likely to benefit from anti-TGFβ therapy. Anti-TGFβ therapy aims to target both the tumor cell and the tumor microenvironment and may well have systemic effects of relevance to tumorigenesis. Extra-tumoral targets include stromal fibroblasts, endothelial and pericyte cells during angiogenesis, and the local and systemic immune systems, all of which can contribute to the pro-oncogenic effects of TGFβ. Many different approaches have been considered, such as interference with ligand synthesis using oligonucleotides, sequestration of extracellular ligand using naturally-occurring TGFβ binding proteins, recombinant proteins or antibodies, targeting activation of latent TGFβ at the cell surface, or signal transduction within the cell. Consideration of which patients might benefit most from anti-TGFβ therapy should include not only tumor responses to TGFβ (which depend on activation of other oncogenic pathways in the cancer cell), but also germline genetic variation between individuals. Ultimately, a deep understanding of the interacting networks of signal pathways that regulate TGFβ outcome in tumor and host cells should allow judicial choice of drugs. This review discusses the progress made in the pre-clinical and clinical testing of TGFβ inhibitors, and discusses considerations of target populations and potential drug regimens.