Therapeutically potent macromolecular drugs have shown great promise for overcoming the limitations of small-molecule anti-cancer drugs. But tumor cell-selective intracellular delivery of the macromolecules remains a major hurdle for their successful clinical application. To overcome this challenge, we engineered a novel genetic fusion protein (F3-Gel) that composed of F3 peptide, a tumorhoming peptide, and gelonin, a plant-derived ribosome-inactivating protein (RIP), and then evaluated its anti-cancer activity in vitro and in vivo. The F3-Gel-encoding gene was synthesized by genetic recombination, and F3-Gel was successfully expressed in E coli. The anti-cancer activity of the produced F3-Gel was evaluated by various in vitro assays, which revealed that F3-Gel maintained equipotent protein synthesis inhibition activity (IC 50 =11 pmol/L) as unmodified gelonin (IC 50 =10 pmol/L). Furthermore, F3-Gel displayed enhanced cellular uptake into cancer cells (U87 MG, HeLa, LnCaP and 9L) than noncancerous cells (293 HEK and SVGp12). Compared with gelonin, F3-Gel exerted significantly higher cytotoxicity against these cancer cells. F3-Gel displayed significantly greater inhibition of protein translation in U87 MG cells: F3-Gel (0.5 μmol/L) was able to reduce the protein level to less than 50%, while gelonin (1 μmol/L) did not affect the intracellular protein level. In a U87 MG xenograft tumor-bearing mouse model, F3-Gel was accumulated in the tumor site at much higher levels and maintained for a prolonged time compared with gelonin. Administration of F3-Gel (0.5, 0.75 mol/kg, iv) caused 36% and 66%, respectively, inhibition of tumor growth in U87 MG xenograft mice, suggesting that it is a promising candidate drug for cancer treatment. Furthermore, this study demonstrates that fusion of F3 peptide to a potent macromolecule could provides an effective method for targeting tumors and eventually could improve their druggability.