Objective To establish the overall diagnostic accuracy of the measurements of vascular endothelial growth factor (VEGF) for malignant ascites. Methods After a systematic review of current studies, sensitivity, specificity, and other measures of the value of ascites concentrations of VEGF in the diagnosis of malignant ascites were pooled by using random effects models. Qualified studies on evaluation of VEGF in diagnosis of malignant ascites in English and Chinese published from January 1990 to December 2009 were retrieved from The Cochrane Library, Cochrane Central Register of Controlled Trials, MEDLINE, EMbase, China National Knowledge Infrastructure (CNKI) databases, WanFang Data, and VIP Information. Two reviewers independently assessed the methodological quality of each study with the tool of QUADAS. Statistical analyses were performed by employing Meta-Disc 1.4 software. Meta-analyses of the reported sensitivity and specificity of each study and Summary Receiver Operating Characteristic (SROC) curve were performed. Results Seven studies met the inclusion criteria for the analysis. After testing the heterogeneity of the included studies, a random effect model was selected to calculate the pool weighted sensitivity and specificity with 95% confidence interval: the sensitivity was 0.81 (95%CI 0.75 to 0.85), the specificity was 0.90 (95%CI 0.86 to 0.94), the DOR was 50.45 (95%CI 28.37 to 89.73), and the AUC of SROC was 0.9507 (SE=0.013 0). The subgroups were analyzed to identify the sources of heterogeneity according to race and agent sources. There was homogeneity among the three studies with agents from Ramp;D company (χ2=0.05, P=0.9750; I2=0.0%), and the AUC of SROC were 0.9675 (SE=0.016 7). Conclusion VEGF has a highly accurate sensitivity and specificity with a b ROC curve, which makes it a new marker to differentiate malignant ascites from the benign.