Therapeutics, based on small interfering RNA (siRNA), have demonstrated tremendous potential for treating cancer. However, issues such as non-specific targeting, premature degradation, and the intrinsic toxicity of the siRNA, have to be solved before they are ready for use in translational medicines. To address these challenges, nanotechnology-based tools might help to shield siRNA and ensure its specific delivery to the target site. Besides playing a crucial role in prostaglandin synthesis, the cyclo-oxygenase-2 (COX-2) enzyme has been reported to mediate carcinogenesis in various types of cancer, including hepatocellular carcinoma (HCC). We encapsulated COX-2-specific siRNA in Bacillus subtilis membrane lipid-based liposomes (subtilosomes) and evaluated their potential in the treatment of diethylnitrosamine (DEN)-induced hepatocellular carcinoma. Our findings suggested that the subtilosome-based formulation was stable, releasing COX-2 siRNA in a sustained manner, and has the potential to abruptly release encapsulated material at acidic pH. The fusogenic property of subtilosomes was revealed by FRET, fluorescence dequenching, content-mixing assay, etc. The subtilosome-based siRNA formulation was successful in inhibiting TNF-α expression in the experimental animals. The apoptosis study indicated that the subtilosomized siRNA inhibits DEN-induced carcinogenesis more effectively than free siRNA. The as-developed formulation also suppressed COX-2 expression, which in turn up-regulated the expression of wild-type p53 and Bax on one hand and down-regulated Bcl-2 expression on the other. The survival data established the increased efficacy of subtilosome-encapsulated COX-2 siRNA against hepatocellular carcinoma.