Objective Hepatocellular carcinoma (HCC), mostly derived from hepatitis or cirrhosisis, is one of the most common types of liver cancer. T-cell mediated immune response elicited by CD/TK double suicide gene has shown a substantial antitumor effect in HCC. Integrin αVβ3 over expresssion has been suggested to regulate the biology behavior of HCC. In this study, we investigated the strategy of incorporating CD/TK double suicide gene and anti-αVβ3 integrin monoclonal antibodies into cationic microbubbles (CMBsαvβ3), and evaluated its killing effect in HCC cells. Methods To improve the transfection efficiency of targeted CD/TK double suicide gene, we adopted cationic microbubbles (CMBs), a cationic delivery agent with enhanced DNA-carrying capacity. The ultrasound and high speed shearing method was used to prepare the non-targeting cationic microbubbles (CMBs). Using the biotin-avidin bridge method, αVβ3 integrin antibody was conjugated to CMBs, and CMBsαvβ3 was generated to specifically target to HepG2 cells. The morphology and physicochemical properties of the CMBsαvβ3 was detected by optical microscope and zeta detector. The conjugation of plasmid and the antibody in CMBsαvβ3 were examined by immunofluorescent microscopy and flow cytometry. The binding capacities of CMBsαvβ3 and CMBs to HCC HepG2 and normal L-02 cells were compared using rosette formation assay. To detect EGFP fluorescence and examine the transfection efficiencies of CMBsαvβ3 and CMBs in HCC cells, fluorescence microscope and contrast-enhanced sonography were adopted. mRNA and protein level of CD/TK gene were detected by RT-PCR and Western blot, respectively. To evaluate the anti-tumor effect of CMBsαvβ3, HCC cells with CMBsαvβ3 were exposed to 5-flurocytosine / ganciclovir (5-FC/GCV). Then, cell cycle distribution after treatment were detected by PI staining and flow cytometry. Apoptotic cells death were detected by optical microscope and assessed by MTT assay and TUNEL-staining assay. Results CMBsαvβ3 had a regular shape and good dispersion. Compared to CMBs, CMBsαvβ3 had more stable concentrations of αVβ3 ligand and pEGFP-KDRP-CD/TK, and CMBsαvβ3 was much sticker to HepG2 HCC cells than normal liver L-02cells. Moreover, after exposed to anti-αVβ3 monoclonal antibody, the adhesion of CMBsαvβ3 to HepG2 cells and L-02 cells were significantly reduced. Also, CMBsαvβ3 demonstrated a substantially higher efficiency in pEGFP-KDRP-CD/TK plasmid transfection in HepG2 cells than CMBs. In addition, CMBsαvβ3 could significantly facilitate 5-FC/GCV-induced cell cycle arrest in S phase. Moreover, treatment of 5-FC/GCV combined with CMBsαvβ3 resulted in a marked apoptotic cell death in HepG2 and SK-Herp-1 HCC cells. In vitro, treatment of 5-FC/GCV combined with CMBsαvβ3 suppresed cell proliferation. In nude mice model, 5-FU + GCV combined with plasmid + CMBsαvβ3were able to significantly suppress tumor volumes. Conclusion Through biotin-avidin mediation system, CMBsαvβ3 were successfully generated to specifically target HCC HepG2 cells. More importantly, CMBsαvβ3 could significantly facilitate 5-FC/GCV-induced cell cycle arrest and apoptotic cell death in HepG2 cells. Our study demonstrated a potential strategy that could be translated clinically to improve liver tumor gene delivery.