Objectives: To assess whether a weaning protocol reduces the mechanical ventilation (MV) duration compared to physician's judgement-based weaning in neurological patients and to determine whether patient consciousness influences this reduction. Methods: A randomised controlled trial was conducted in a neurological intensive care unit (NCU) of a tertiary hospital; 144 patients requiring MV for more than 24 hours were randomly allocated to protocol-directed (intervention) (n=71) or physician-directed (control) group (n=73). Results: The intervention group displayed a significantly shorter median weaning time than the control group (2.00 vs 5.07 days, P<0.05). The median MV duration tended to be shorter in the intervention group (10.8 vs 14.2 days, P=0.106). The median length of NCU stay was 19.0 and 26.1 days in the intervention and control groups, respectively (P=0.063). The median NCU cost was 9.26 X 10 (4) and 12.24 X 10 (4) (sic)in the intervention and control groups, respectively (P=0.059). The unsuccessful weaning, ventilator-associated pneumonia (VAP) and mortality rates were similar between the groups. Among conscious patients, the median weaning time (2.00 vs 7.00 days, P<0.05) and the median MV duration (8.8 vs 18.0 days, P=0.017) were significantly reduced in the intervention group. Among unconscious patients, the intervention group displayed a reduced median weaning time (1.00 vs 3.10 days, P<0.05), but not median MV duration (11.6 vs 11.1 days, P=0.702), compared to the control group. Conclusion: Protocol-directed weaning reduces weaning time, MV duration, length of NCU stay and NCU cost in neurological patients, and these effects are more significant in conscious patients than in unconscious patients.