Objectives Measurements of stroke volume variation for volume management in mechanically ventilated patients are influenced by various factors, such as tidal volume, respiratory rate, and chest/lung compliance. However, research regarding the effect of positive end-expiratory pressure on stroke volume variation is limited. Methods Patients were divided into responder and nonresponder groups according to the prediction of fluid response by the passive leg raising test and hemodynamic parameters, including stroke volume variation, measured in all patients at the following ventilator settings: (1) conventional ventilation (C), tidal volume 10 mL·kg−1 with positive end-expiratory pressure settings of 0 (C0), 5 (C5), and 10 cmH2O (C10) and (2) lung protective ventilation (P), tidal volume 6 mL·kg−1 with positive end-expiratory pressure settings of 0 (P0), 5 (P5), and 10 cmH2O (P10). Results Regardless of ventilator setting, stroke volume variation in the responder group had an increasing trend as increased positive end-expiratory pressure level and was significantly higher than in the nonresponder group at each positive end-expiratory pressure level. The area under the curve was (1) 0.899 at C0, 0.942 at C5, and 0.985 at C10; and (2) 0.901 at P0, 0.932 at P5, and 0.947 at P10. Optimal threshold values given by receiver operating characteristic curve analysis were (1) 13.5%, 13.5%, and 14.5%; and (2) 13.5%, 13.5%, and 14.5%, respectively. Conclusions The threshold value of stroke volume variation in predicting fluid responsiveness may change when positive end-expiratory pressure 10 cmH2O is applied. This must be considered when stroke volume variation is used to detect the fluid responsiveness to prevent volume overload in this mechanical ventilation setting.