We conducted a six-month prospective interventional crossover study examining a computerized diabetes monitoring system (DMS) that conveyed dietary information. The objectives were to compare glycaemic control between intervention and control periods, and to assess patients' acceptance of the DMS. Nineteen patients were randomized into two groups, each using the DMS for three months and serving as the control group for another three months. The patients recorded information about their meal portions and blood glucose readings in a hand-held electronic diary. After transmitting the data to the DMS through a telephone modem, the patients received immediate feedback about the carbohydrate, protein and fat content of the meal, as well as the calorie content. A significant improvement in glycaemic control was achieved during intervention compared with control periods (mean HbA1C reduction of 0.825%). The DMS was also highly acceptable: 95% patients found it easy to operate while 63% found it useful. The DMS was thus a feasible model of telemedicine in diabetes care and a larger study is warranted to examine its cost-effectiveness.