Objective: To systematically evaluate the diagnostic value of circulating tumor cells (CTCs) in patients with pancreatic cancer through Meta-analysis. Methods: The studies regarding detection of CTCs in pancreatic cancer patients and non-pancreatic cancer subjects were collected through searching several national and international databases. The retrieval time was from inception of the database to 31 August 2016. After screening for inclusion criteria, data extraction and quality assessment, Meta-analysis was performed by the Metadisc 1.4 and Stata 12.0 software. Results: A total of 19 articles were selected, involving 693 pancreatic cancer patients and 406 non-pancreatic controls. Pooling the data using the random effect model showed that the pooled sensitivity and specificity of the CTCs for detection of pancreatic cancer was 0.67 (95% CI=0.63–0.60) and 0.94 (95% CI=0.91–0.96), the pooled diagnostic odds ratio was 50.47 (95% CI=20.13–126.55), and the pooled positive likelihood ratio and negative likelihood ratio was 11.15 (95% CI=5.42–22.95) and 0.36 (95% CI=0.28–0.45), respectively. The area under the curve of the summary receiver operating characteristic curve was 0.93 with a Q* value of 0.03. Diagnostic Test Deek’s funnel plot for diagnostic test demonstrated the presence of publication bias. Conclusion: CTCs detection is incompetent to be used dependently as an early diagnosis indicator for pancreatic cancer, while it may probably serve as a supplementary index for diagnosis of pancreatic cancer.