Background: This study was performed to assess the association of TLR4 gene 2026A/G (rs1927914), 896A/G (rs4986790), and 1196C/T (rs4986791) polymorphisms and cancer susceptibility based on published case-control studies. Methods: Web of Science, PubMed, Embase, CBM, WanFang Data, CNKI, and VIP database were used for article retrieving. Then, these articles were screened according to the study inclusion and exclusion criteria. The data was extracted, and the study quality was evaluated according to the principle of Newcastle-Ottawa Scale. Meta-analysis was performed by RevMan 5.4 and Stata MP-17 software. Trial sequential analysis was performed by TSA 0.9.5.10 Beta software. Results: Eighty-seven case-control studies including 25,969 cases and 32,119 controls were included in the meta-analysis. The diseases involved in case groups include prostate cancer, lung cancer, gastric cancer, hepatocellular carcinoma, colorectal cancer, etc. A versus G model of rs1927914, A versus G model of rs4986790 and C versus T model of rs4986791 showed that odds ratio (OR) = 1.08, OR = 0.85, and OR = 0.74 respectively. All the 3 comparisons were statistically significant. Sensitivity analysis showed that the results were stable. Publication bias analysis and trial sequential analysis showed that no significant publication bias was found in the results of the meta-analysis, and the probability of false positives was small. Conclusion: People with A allele of rs1927914, G allele of rs4986790, or T allele of rs4986791 have higher risks of cancer. The results of meta-analysis are stable and have less probability of false positives.