Fluid management during thoracic anesthesia remains as a challenge for the anesthesiologists. The "safe zone" between volume overload (risk of pulmonary edema) and hypovolemia (potential risk of renal failure) is hard to determine and narrow. Relationship between perioperative fluid administration and postoperative lung injury has been clearly demonstrated but lung injury can even occur after the most restrictive management. Multiple hit hypothesis and endothelial glycocalyx in addition to revised Starling equation can help us understand this dilemma. Although a "liberal" fluid strategy is out of question for thoracic surgery, evidence on "restrictive" and "goal-directed" fluid strategies are missing. New monitors with dynamic parameters can help to evaluate if the patient will respond to a fluid challenge. However, "volume responsiveness" should not be equated with volume deficiency, especially in a patient with thoracic epidural catheter. Fluid type (crystalloids or colloids) is also another unclear point, although balanced solutions should be preferred for crystalloids. Minimal discontinuation of oral hydration, early feeding and mobilization should be encouraged. This review summarizes current evidence on the topic and highlights unanswered questions.