The use of elective tracheostomy in major head and neck surgery is well established, although practice varies between units. There is no published method that reliably predicts the need for tracheostomy. This paper describes the development of a surgical scoring system designed to achieve that aim. The system was devised using data obtained retrospectively from 148 consecutive major head and neck procedures. These procedures were grouped according to the airway management plan in place at the end of the procedure: elective extubation (group E, 52 procedures, 50 patients); elective overnight ventilation via an endotracheal tube (group ETT, 55 procedures, 52 patients); and elective overnight ventilation via a tracheostomy (group T, 41 procedures, 41 patients). 8 patients from group ETT required a late tracheostomy for either medical or surgical indications. Using statistical methods, a threshold score was defined above which the high risk of upper airway obstruction should prompt consideration of an elective tracheostomy.