Therapeutic alliance is a well-recognized predictor of patient outcomes within psychological therapy. It has not been applied to obesity interventions, and Bordin's theoretical framework shows particular relevance to the management of obesity in primary health care. This cross-sectional study of a weight management programme in general practice aimed to determine if therapeutic alliance was associated with patient outcomes. The Working Alliance Inventory short revised version (WAI-SR) was administered to 23 patients and 11 general practitioners (GPs) at the end of a 6-month weight management programme. Use of the WAI-SR indicated that the strength of therapeutic alliance varied between different patient-GP relationships in this pilot intervention. A robust therapeutic alliance was strongly associated with patient engagement in the weight management programme indicated by number of appointments. It was also associated with some general health and quality of life outcomes. These are promising results that require confirmation with larger studies in primary health care. The measurement of therapeutic alliance using the WAI-SR may predict patient attendance and outcomes in obesity interventions in primary healthcare settings.