Several studies have shown that the placement of a face mask on a horse can have effects on ventilation, gas exchange and the cardiovascular system during exercise. The aim of the present study was to determine if airflow and ventilation measured with the same ultrasonic flowmeters were different during exercise between horses wearing half-(HM) and full-face (FM) masks. Five clinically healthy Thoroughbred horses with no history of respiratory disease were studied in an unbalanced crossover design. They were exercised on a treadmill at speeds between 1.7 and 11 m s−1 on a 3° incline wearing both masks. The following variables were recorded: peak inspired (PIF) and peak expired flow rates (PEF), inspiratory tidal volume (VT), respiratory rate (fR), inspiratory minute ventilation (VE), inspiratory time, (TI), expiratory time (TE), total breath time (TT), end tidal oxygen (ETO2), end tidal carbon dioxide (ETCO2) and heart rate (HR). A mask by speed of exercise interaction term was not significant for any of the models. The PEF (mean difference 12.9 l s−1; lower and upper 95% CI 7.6 and 18.2 l s−1, respectively; P < 0.0001) and ETO2 (mean difference 0.77%; lower and upper 95% CI 0.48 and 1.00%, respectively; P < 0.0001) were significantly greater and ETCO2 was significantly lower (mean difference −1.3%; lower and upper 95% CI −2.0 and 0.7%, respectively; P < 0.0001) with the FM compared with the HM. There was also a trend for inspired VE to be higher with the FM compared with the HM (mean difference 102 l min−1; lower and upper 95% CI 26 and 178 l min−1, respectively; non-significant). We conclude that the HM may impair ventilation in the horse during exercise compared with the FM, despite the latter having a greater deadspace.