We conducted a crossover randomised study to evaluate the performance of a novel optical stylet, the InnoScope, for tracheal intubation in simulated normal and difficult airways. Twenty-five anaesthetists attempted tracheal intubation on a SimMan 3G simulator using the InnoScope first followed by the Macintosh laryngoscope or vice versa. Three airway scenarios were tested: (1) normal airway; (2) difficult airway with swollen pharynx; and (3) limited neck movement. In each scenario, the laryngeal view, duration of and success rate for tracheal intubation were recorded. Compared with the Macintosh laryngoscope, the use of InnoScope increased the percentage of glottic opening seen by 17% in normal airway, 23% in the difficult airway and 32% with limited neck movement, p < 0.01. Despite this better laryngeal view, successful tracheal intubation achieved with the InnoScope (88.0%) was lower than that for the Macintosh laryngoscope (98.7%), p = 0.008. Using the InnoScope, tracheal intubation during the first attempt was only successful in 48% of cases with difficult airway. In this scenario, the median (interquartile range [range]) duration of tracheal intubation was significantly longer with the InnoScope (53 (20–100 [15–120]) s) compared with the Macintosh laryngoscope (27 (20–62 [15–120]) s), p = 0.01. We conclude that an improved laryngeal view with the use of the InnoScope did not translate into better conditions for tracheal intubation.