The structure and function of a 27-a.a. fragment of the N-terminal sequence of human endostatin (ES-Zn) were compared to those of the mutant peptide (ES-SSZn) obtained by adding Cys-Pro-Ala to the endostatin N-terminus and substituting Asn16 for Cys ensuring formation of a disulfide bond. Structural comparison of ES-Zn and ES-SSZn by far-UV circular dichroism (CD), intrinsic fluorescence, and molecular dynamics simulation methods revealed significant structural perturbations in ES-SSZn, such as elimination of the β-sheet conformer, modification of the N-terminal loop structure, and reorganization of dynamic properties of the entire peptide backbone. ES-SSZn was approximately 2 and 3 times less efficient than ES-Zn and the full-length human endostatin, respectively, in the induction of caspase-3-dependent apoptosis in human umbilical vein endothelial cells (HUVECs) in vitro (p < 0.05). In contrast, treatment of metastatic 4T1 breast tumors in mice with ES-Zn and ES-SSZn (5 mg/kg body weight daily) for 14 days resulted in similar regression of tumor size, comparable downregulation of angiogenesis (CD31 and CD34) and cell proliferation (Ki67), and therefore, the same extent of apoptosis induction (TUNEL, p53, and Bcl-2) for both peptides (as compared to the untreated controls). Western blot analysis of HUVEC and 4T1 tumor lysates revealed the same levels of suppression of key signaling mediators Akt and ERK1/2 by ES-Zn and ES-SSZn. Contrary to the earlier studies, our results showed that the function of the 1-27 endostatin fragment is independent of its overall structure. Stabilization of the N-terminal loop structure by the disulfide bond incorporation causes relief from structural deviations.