Objective To investigate the associations of polymorphisms at the rs3819122 and rs12456284 loci in the SMAD4 gene with susceptibility of hepatocellular carcinoma (HCC) and patients' survival prognosis. Methods A case-control study was conducted, enrolling 202 patients with HCC (case group) and 176 healthy controls (control group), and their clinical data were collected. Matrix-assisted laser desorption/ionization time-of-flight mass spectrometry was used to detect polymorphisms at the rs3819122 and rs12456284 loci of the SMAD4 gene in all participants. The Hardy-Weinberg genetic equilibrium for genotype distribution in each group was assessed using the goodness-of-fit χ² test; unconditional logistic regression model was used to analyze the associations between genotypes and HCC susceptibility; the Kaplan-Meier method and Cox proportional hazards regression model was used to analyze the associations between genotypes and the survival prognosis of HCC patients. Results No statistically significant difference was found between the case group and the control group in terms of age, alcohol consumption, or smoking status (all P>0.05); however, statistically significant differences were observed between the two groups regarding gender and hepatitis B virus (HBV) infection status (all P<0.05). The rs3819122 locus of the SMAD4 gene exhibited 3 genotypes: AA, CA, and CC; the rs12456284 locus also exhibited 3 genotypes: AA, GA, and GG; results of the Hardy-Weinberg equilibrium test showed that the distribution of genotype frequencies in both groups conformed to the law of genetic equilibrium (all P>0.05). Unconditional logistic regression analysis revealed: (1) at the rs3819122 locus, individuals carrying the CC genotype and the CA+CC genotype had a 0.228-fold (95%CI: 0.117-0.756, P=0.011) and a 0.417-fold (95%CI: 0.199-0.873, P=0.020) risk of HCC, respectively, compared with those carrying the AA genotype; (2) at the rs12456284 locus, no statistically significant association was found between any genotype and HCC susceptibility (all P>0.05). Stratified analysis using gender and HBV infection as stratification variables showed that the protective effects of the CC genotype and CA+CC genotype at the rs3819122 locus were statistically significant only in males and in HBV-negative individuals (all P<0.05); the GA+GG genotype at the rs12456284 locus was associated with reduced HCC susceptibility only in the HBV-negative population (P<0.05). Kaplan-Meier survival analysis showed no statistically significant difference in the survival status among HCC patients with different genotypes at the rs3819122 and rs12456284 loci (all P>0.05). Cox proportional hazards regression analysis indicated no statistically significant association between polymorphisms at the rs3819122 and rs12456284 loci and the risk of death in HCC patients (all P>0.05). Conclusion Polymorphism at the rs3819122 locus of the SMAD4 gene is associated with HCC susceptibility, the CC genotype and CA+CC genotype may reduce the risk of HCC, and their protective effects are observed only in males and HBV-negative individuals; the GA+GG genotype at the rs12456284 locus is associated with reduced HCC susceptibility only in the HBV-negative population. Furthermore, this study found no association between polymorphisms at the rs3819122 and rs12456284 loci and the survival prognosis of patients with HCC.