Like most embryonic tissues, tumors have the ability to build up their own blood vessel networks. However, the architecture of tumor vessels is fundamentally different from that found in healthy tissues. Tumor vessels are usually irregular, heterogeneous, leaky, and poorly associated with mural cells. Endothelial cells in tumor vessels are also disorganized and express imbalanced surface molecules. These unusual features may provide some molecular and structural basis for selective inhibition or even destruction of tumor vessels by angiogenesis inhibitors. In animal tumor models, several angiogenesis inhibitors seem to inhibit tumor angiogenesis specifically without obvious effects on the normal vasculature. As a result, these inhibitors produced potent antitumor effects in mice. Excited by these preclinical studies, more than 60 angiogenesis inhibitors are being evaluated for their anticancer effects in human patients. Although the ultimate outcome of antiangiogenic clinical trials remains to be seen, several early observations have reported some disappointing results. These early clinical data have raised several important questions. Can we cure human cancers with angiogenesis inhibitors? Have we found the ideal angiogenesis inhibitors for therapy? What is the difference between angiogenesis in an implanted mouse tumor and in a spontaneous human tumor? What are the molecular mechanisms of these angiogenesis inhibitors? Should angiogenesis inhibitors be used alone or in combinations with other existing anticancer drugs? In this review, we will discuss these important issues in relation to ongoing antiangiogenic clinical trials.