Background: The effect of breathing modification techniques on asthma symptoms and objective disease control is uncertain. Methods: A prospective, parallel group, single-blind, randomised controlled trial comparing breathing training with asthma education (to control for non-specific effects of clinician attention) was performed. Subjects with asthma with impaired health status managed in primary care were randomised to receive three sessions of either physiotherapist-supervised breathing training (n = 94) or asthma nurse-delivered asthma education (n = 89). The main outcome was Asthma Quality of Life Questionnaire (AQLQ) score, with secondary outcomes including spirometry, bronchial hyper-responsiveness, exhaled nitric oxide, induced sputum eosinophil count and Asthma Control Questionnaire (ACQ), Hospital Anxiety and Depression (HAD) and hyperventilation (Nijmegen) questionnaire scores. Results: One month after the intervention there were similar improvements in AQLQ scores from baseline in both groups but at 6 months there was a significant between-group difference favouring breathing training (0.38 units, 95% CI 0.08 to 0.68). At the 6-month assessment there were significant between-group differences favouring breathing training in HAD anxiety (1.1, 95% CI 0.2 to 1.9), HAD depression (0.8, 95% CI 0.1 to 1.4) and Nijmegen (3.2, 95% CI 1.0 to 5.4) scores, with trends to improved ACQ (0.2, 95% CI 0.0 to 0.4). No significant between-group differences were seen at 1 month. Breathing training was not associated with significant changes in airways physiology, inflammation or hyper-responsiveness. Conclusion: Breathing training resulted in improvements in asthma-specific health status and other patient-centred measures but not in asthma pathophysiology. Such exercises may help patients whose quality of life is impaired by asthma, but they are unlikely to reduce the need for anti-inflammatory medication.