Weaning from mechanical ventilation is complicated in patients with severe brain injury, but recurrent stimulation by familiar voices and commands to breathe in and out during critical weaning periods may improve patient outcomes. This study aimed to assess the feasibility, safety and efficacy of audio recordings of patients' relatives to support weaning from mechanical ventilation. VOICE-WEANING II (Voices of patients' relatives to support weaning from mechanical ventilation) was a randomized (1:1), sham-controlled clinical trial. Patients aged 18 years and older with controlled mechanical ventilation ≥ 48 h due to neurological disease were included. Patients received either audio recordings or sham control with muted audio recordings for 10 min, three times per day from initiation of assisted mechanical ventilation. The primary outcome was rate of weaning failure. Secondary outcomes included duration of controlled ventilation, rate of tracheostomy, delirium and all-cause mortality at 90 days. Brain activity was assessed using spectral density analysis of continuous electroencephalogram monitoring. Fourty-five participants were randomized (25 males/20 females, median age 60 years). Of those, 22 patients received audio recordings (48.9%) and 23 (51.1%) sham control. Rate of weaning failure was 52.4% in the intervention group and 63.6% in the control group (adjusted difference -9.5; 95% confidence interval, -38.8 to 19.9; P = 0.50). Duration of controlled mechanical ventilation was significantly reduced in the treatment group (adjusted difference -19.4 h; 95% confidence interval, -37.4 to -1.5 h; P = 0.03). The intervention was feasible and safe. Brain activity was increased in response to treatment and pronounced in right fronto-central brain regions. Although audio recordings of patients' relatives did not significantly reduce weaning failure, the duration of controlled mechanical ventilation was significantly reduced and brain activity increased suggesting an immediate treatment response. These trial results seem to indicate a therapeutic effect of audio recordings of patients' relatives for weaning from mechanical ventilation.