Morphine for dyspnea control usually arouses ethical controversy in terminal cancer care. This study prospectively assessed the use of morphine for dyspnea control in terminal cancer patients in terms of two characteristics: the extent to which medical staff, family, and patients found morphine to be ethically acceptable and efficacious, and the influence of morphine on survival. One hundred and thirty-six palliative care patients meeting specific eligibility criteria were enrolled. A structured data collection form was used daily to evaluate clinical conditions, which were analyzed at the time of admission and 48 h before death. Sixty-six (48.6%) of the 136 patients had dyspnea on admission. The intensity was mild in 14.0% and moderate or severe in 34.6%. The intensity of dyspnea became worse 48 h before death (4.29+/-2.55 vs. 4.94+/-2.60, P < 0.001, range 0-10). Twenty-seven (40.9%) of 66 patients with dyspnea received morphine on admission for the control of dyspnea; the routes of administration were oral (59.3%) and subcutaneous (40.7%). Fewer than two-thirds (59.3%) of the patients were given morphine with the consent of both patient and family. More than one-third (40.7%) on admission and about one-half (52.8%) in the 48 h before death had the consent of family alone. Positive ethical acceptability and satisfaction with using morphine for dyspnea control were found in both medical staff and family in this study. Multiple Cox regression analysis showed that using morphine for dyspnea, both on admission and in the 48 h before death, did not significantly influence the patients' survival (HR: 0.015, 95% CI: 0.00-4.23; HR: 1.76, 95% CI: 0.73-4.24). In this population, the use of morphine for dyspnea control in the terminal phase of cancer was effective and ethically validated in the study. Research efforts to find the most appropriate route and dosage of morphine for dyspnea, based on the patient's situation, remain worthwhile.