This article serves to evaluate the association of polymorphisms of mismatch repair genes (hMLH1 and hMSH2) with breast cancer (BC) susceptibility through a meta-analysis. Our methods involved extensive research in Chinese and English databases that examined the association of hMLH1 and hMSH2 polymorphisms with susceptibility to BC, strictly abiding by established inclusion and exclusion criteria. Software Stata 12.0 was used for statistical data analysis. A total of 12 studies were available for meta-analysis, published between 2014 and 2017, of which respectively 9 studies explored the association of hMLH1 (rs1799977 A > G and rs63750447 T > A) and 3 studies explored the association of hMSH2 (rs4987188 [Gly322Asp] and rs17217772 [Asn127Ser]) with patients' susceptibility to BC. The results showed that both the rs1799977 A > G polymorphism GA + GG genotype (especially in the Caucasian population) and the rs63750447 T > A polymorphism TA + AA genotype in the hMLH1 gene increased patients' susceptibility to BC. The genotype detection method was selected as a target for subgroup analysis. According to studies where MassARRAY assay was conducted, the rs1799977 A > G polymorphism was correlated with BC susceptibility in the dominant model, while rs4987188 (Gly322Asp) and rs17217772 (Asn127Ser) of the hMSH2 gene presented no observable correlation with the risk for BC. Both the rs1799977 A > G and rs63750447 T > A polymorphisms in the hMLH1 gene showed a significant association with a markedly increased risk for BC, while rs4987188 (Gly322Asp) and rs17217772 (Asn127Ser) of the hMSH2 gene were not clearly correlated with BC susceptibility.