The risk of testicular cancer (TC) is markedly increased in subjects with androgen insensitivity, and previous studies have proposed that GGN and CAG repeats in androgen receptors ( AR) could be related to the risk of TC. To evaluate the association between the length of GGN and CAG repeats in AR and TC, a meta-analysis involving 3255 TC cases and 2804 controls was performed. The results suggested that long GGN repeats are associated with an increased risk of TC compared with those < 23 [odds ratio (OR) = 1.22, 95% confidence interval (CI) = 1.05–1.41]; similarly, a subgroup analysis revealed that this association occurred in studies with case sizes > 200, and in the mid-latitude, and seminoma subgroups. The subgroup analysis based on populations, high-latitude, and seminomas/non-seminomas suggested that AR CAG repeat polymorphisms with > 25 and < 21 + > 25 repeats might confer a protective effect to the patients with TC (in the high-latitude subgroup analysis, for > 25 vs. 21–25: OR = 0.54, 95% CI = 0.41–0.70). In contrast, an increased risk of TC was observed for AR CAG repeat polymorphisms with > 25 and < 21 + > 25 repeats in the mid-latitude subgroup (for > 25 vs. 21–25: OR = 1.65, 95% CI = 1.09–2.50). In addition, no associations between the remaining subgroups and male infertility were observed. In short, this meta-analysis suggested that AR GGN and CAG repeat polymorphisms may be involved in the etiology of TC.