Clostridium difficile infection (CDI) causes nosocomial antibiotic-associated diarrhea and colitis in the developed world. Two potent cytotoxins, toxin A (TcdA) and toxin B (TcdB) are the virulence factors of this disease and can be a good vaccine candidate against CDI. In the present study, we genetically engineered Lactococcus lactis to express the nontoxic, recombinant fragments derived from TcdA and TcdB C-terminal receptor binding domains (Tcd-AC and Tcd-BC) as an oral vaccine candidate. The immunogenicity of the genetically engineered L. lactis oral vaccine delivery system (animal groups LAC and LBC or the combination of both, LACBC) was compared with the recombinant TcdA and TcdB C-terminal receptor binding domain proteins (animal groups PAC and PBC or the combination of both, PACBC), which were expressed and purified from E. coli. After the C. difficile challenge, the control groups received PBS or engineered L. lactis with empty vector, showed severe diarrhea symptoms and died within 2-3 days. However, both the oral vaccine and recombinant protein vaccine groups had significantly lower mortalities, body weight decreases and histopathologic lesions than the control sham-vaccine groups (p<0.05) except group LBC which only had a 31% survival rate after the challenge. The data of post infection survival showed that an average of 86% of animals survived in groups PAC and PACBC, 75% of animals survived in group LACBC, and 65% of animals survived in group LAC. All of the vaccinated animals produced higher titers of both IgG and IgA than the control groups (p<0.05), and the antibodies were able to neutralize the cytopathic effect of toxins in vitro. The results of this study indicate that there is a potential to use L. lactis as a delivery system to develop a cost effective oral vaccine against CDI.