Objective: To examine whether sperm DNA fragmentation has the predictive value for ART outcomes of IVF/ICSI.Methods: An exhaustive electronic literature search was conducted on PubMed, Embase and CNKI databases from database inception to June 2015.The basis of quality assessment for studies of diagnostic accuracy (QUADAS) criteria was used to assess quality of the studies.This bivariate meta-analysis was performed in a random-effects model using sensitivity, specificity, summary receiver operator characteristic (SROC) curves, area under the curve (AUC) , and 95%Cls as effect measures.Heterogeneity was measured using the CochransQ test.Publication bias and subgroup analysis were also conducted through Stata 12.0 software.Results: A total of 6 studies comprising 13 individuals were eligible to analysis outcomes of assisted reproduction treatment(ART).The bivariate random effects meta-analysis showed that an AUC value was 0.6844 (sensitivity 0.53[95% Cl 0.49 to 0.57];specificity 0.58 [95% Cl 0.54 to 0.61])for sperm DNA fragmentation assay in predicting ART outcomes after IVF/ICSl.Notably, the subgroup analysis suggested that ICSl treatment of sperm DNA fragmentation assay (AUC 0.6974;sensitivity 0.67;specificity 0.63) seemed to obtain higher accuracy than IVF treatment (AUC 0.6589;sensitivity 0.35;specificity 0.52) in predicting ART outcomes.There was a significant heterogeneity among studies for subgroup analysis.Conclusions: The results indicate that assays detecting sperm DNA fragmentation confers a modest prognostic value for predicting outcome in infertile couples undergoing IVF/ICSI treatments.The data also suggest that sperm DNA fragmentation should be recommended to those suffering from recurrent failure to achieve pregnancy.Selection of sperm without DNA damage for ART may improve the clinical outcomes of IVF/ICSI.