Background: Mannitol is widely used to reduce brain tissue swelling and improve brain relaxation during neurosurgery. However, the optimal dosage for patients with midline shift undergoing supratentorial tumor resection remains unclear. Methods: In this randomized, controlled double-blinded study, 204 patients with preoperative midline shift who underwent elective supratentorial brain tumor surgery were equally allocated to receive placebo or 0.7, 1.0, or 1.4 g/kg mannitol infusion. The primary outcome was the proportion of satisfactory brain relaxation. Results: Demographics and baseline characteristics were similar among the 4 groups. Trend analysis showed that mannitol infusion increased satisfactory brain relaxation (P<0.0001), relaxed dural tension (P<0.0001) and adequate surgical exposure (P<0.0001), and decreased the requirement for rescue therapy for brain swelling (P<0.0005), all in a dose-dependent manner. Tumor size (odds ratio [OR]: 0.99 per 1mm3, 95% confidence interval [CI]: 0.989-0.998, P=0.004), peritumoral edema classification (OR: 0.60, 95% CI: 0.37-0.97; P=0.038) as well as mannitol dose (OR: 2.81, 95% CI: 1.97-4.02, P<0.0001) were significantly associated with satisfactory brain relaxation. An increased risk of moderate to severe postoperative cerebral edema was found in the group receiving 1.4 g/kg mannitol (P=0.025) in a dose-dependent manner (P=0.018). Conclusions: An optimal mannitol infusion dosage of 1.0 g/kg is recommended to improve brain relaxation with lower risk of moderate to severe postoperative cerebral edema in patients with midline shift undergoing supratentorial tumor resections. The effect of mannitol on brain relaxation is affected by tumor size and severity of peritumoral edema, rather than by midline shift. © 2019 Wolters Kluwer Health, Inc.