According to the recent research available from the American Cancer Society, cancer is the second leading cause of death in the USA. Many kinds of chemotherapeutic agents have been shown to be effective against tumor cells. Unfortunately, most of them also kill normal cells. Because native antibodies and antibody fragments find it hard to penetrate solid tumors, most approaches have focused on developing antibody mimetics. Therefore, we wanted to design more reasonable antibody mimetics, containing three regions, for targeting glioma and small cell lung cancer, and supposed that these mimetic-toxin immunotoxins could specifically kill targeting tumors. VLCDR1 and VHCDR3, together with five amino acid residues on both sides of the complementarity-determining regions, through a cognate framework region (VHFR2) yielded mimetics of BT32/A6 (United States Patent number: 5639863) and TB2A36C3 (United States Patent number: 5744585). When genetically fused, the mimetic with the first 388 amino acid residues of diphtheria toxin, Escherichia coli strain, BL21 (ED3), was used for expression of soluble immunotoxins DT-BT and DT-TB. The killing activities of our fusion proteins were examined in vitro. DT-BT and DT-TB could not kill Raji up to the maximal concentration tested (10⁻⁷ mol/l) in vitro. In contrast, at concentrations ≥ 10⁻⁸ mol/l, DT-BT and DT-TB killed 90% of U-87 MG cells and small cell lung cancer in vitro, respectively. The killing activities of our fusion proteins hinted that 5aa-VLCDR1-5aa-VHFR2-5aa-VHCDR3-5aa mimetics maintained the synergic interactions and high affinity of parent antibodies. Such mimetics may hold promise for markedly improving cancer diagnostics and therapeutics, especially as the small size of the mimetics may confer other useful properties, such as enhanced intracellular delivery.