Pancreatic cancer remains a challenging disease with an overall cumulative 5-year survival rate around 6%. Though significant progress has been made in the availability of diagnostic techniques and treatment strategies, pancreatic cancer remains a disease of high mortality rate. Therefore, there is an urgent need for a better understanding of the molecular mechanisms that governs the oncogenesis and metastasis process of pancreatic cancer. In the present study, by using the Cancer Genome Atlas (TCGA) dataset analysis, we demonstrated that sorting nexin 6 (SNX6) serves as a biomarker for predicting prognosis of pancreatic cancer. In vitro studies demonstrated that silencing of SNX6 expression reduced cell proliferation, colony formation, invasion, and metastasis. Higher level of SNX6 helps maintain the mesenchymal properties, which renders migration and invasive capacities to pancreatic cancer cells. Moreover, in the process of TGF-β-induced epithelial to mesenchymal transition (EMT), the expression level of SNX6 was increased, and silencing of SNX6 expression could inhibit the TGF-β-induced EMT program. These results collectively uncovered a novel predictive marker for pancreatic cancer and provided the possible underlying molecular mechanism.