We studied 20 unselected patients admitted to our Intensive Care Unit (ICU) suffering from acute respiratory failure (ARF), who needed mechanical ventilatory support. In all of them we followed a prospective protocol to investigate the value of mouth occlusion pressure (P0.1) as an indicator for weaning. Fifty-two tests were classified into three groups: a need to be reconnected to mechanical ventilation (MV), stable on intermittent mandatory ventilation (IMV), or spontaneous breathing on a T-tube (TT). The results showed that at increased values of P0.1 there were more difficulties in weaning patients from MV. Seventy-eight percent (78%) of the occasions where weaning was successful, values of P0.1 were less than or equal to 4.2 cm H2O, in chronic or non-chronic patients. Eighty-nine percent (89%) of the times when P0.1 values were higher than 4.2 cm H2O the same patients required ventilatory support, total (MV) or partial (IMV). These differences were statistically significant (p less than 0.01). We conclude that the P0.1 is an easily obtained non-invasive parameter, that can contribute along with other more conventional measurements to a superior indication for weaning.