Ｏｂｊｅｃｔｉｖｅ　To evaluate the ability of pleth variability index (PVI) in predicting fluid responsiveness in the gastrointestinal tract surgery.　Methods　Fifty patients were enclosed in this study following anesthesia induction. PVI was continuously displayed by the Masimo Radical 7. All patients were also monitored with Vigileo/FloTrac system. Haemo-dynamic data such as cardiac index(CI), stroke volume index(SVI), stroke volume variability(SVV), MAP, HR, CVP, PVI, perfusion index(PI) were recorded before and after volume expansion(voluven injection, 7 ml/kg). Positive response was defined as an increase in SVI≥10% (△SVI≥10%).　Results　Forty-eight patients finished the trial. Twenty-eight patients were regarded as positive responders and twenty patients were regarded as negative responders according to positive response criteria. SVV and PVI were significantly higher in the positive responders (13.8±1.2)% and (15.5±4.2)% than those in negative responders (9.9±3.2)% and (11.3±3.2)% respectively (P<0.05). The SVV diagnostic threshold is 10.5%. The volume expansion was able to discriminate the positive responder from the negative responders with a sensitivity of 80.0%. The specificity is 100%. The diagnostic threshold of PVI was 13.5%, the sensitivity is 65.0% and the specificity is 75.0%. There was relativity between PVI before volume expansion and change in SVI after volume expansion is r=0.420(P<0.05). Similar relativity of SVV is r=0.639(P<0.01).　Conclusions　PVI as a new dynamic index can predict fluid responsiveness non-invasively in the gastrointestinal tract surgery during general anesthesia.