In humans and animals, attention is considered to be underlying a variety of cognitive processes, such as learning or memory. Attention is usually described as the ability to process selectively one aspect of the environment over others and in relationship with higher performances in human daily tasks (e.g. at school). The rare studies on animal attention are limited because the paradigms used were adapted from human attention studies; they involve extensive training and imply a laboratory context that is not adapted to field studies. We chose to characterize horses’ attention by designing a novel visual attention test (VAT) that does not require extensive training and that is easy to apply in the field. This test was inspired by an ethological approach based on spontaneous attention behaviour towards a visual moving stimulus without involving operant conditioning: it consisted in projecting for 5 minutes a green light from a laser pointer on the stall door of the horse (with repeated circular clockwise and 50-cm long vertical and horizontal movements). The study was conducted on seventeen horses, including 7 females and 10 geldings, aged from 7 to 12 years ( ±ES = 8.1±1.6), from French Saddlebred (n=13) and Anglo- Arabian (n=4) breeds and were ridden for either jumping (n=10) or eventing (n=7) competitions. Each horse was tested with the VAT once a day for two consecutive days (i.e. day 1 and day 2) and its competition performance index was collected. By measuring all horses’ gazes towards the stimulus, our VAT revealed different patterns of attention that can indicate a horse’s attention level: overall visual attention when the horse merely gazed at the stimulus, and “fixed” attention characterized by fixity and orientation of at least the visual and auditory organs towards the stimulus. Our results also revealed that the sequences of attention were very short (3.7 seconds on average) and fragmented, suggesting frequent refocusing of attention. Overall, the more horses’ attention was fragmented, the higher the index of performance was (Spearman’s correlation test, N=17, rs=0.47, p=0.05) in particular for eventing horses (Spearman’s correlation test, N=7, rs=0.73, p=0.05). This novel test seems a promising tool to study the attentional characteristics and skills in horses. Horses’ attention characteristics such as attention fragmentation during the VAT can be predictive of equine performances in competition. Previous studies revealed altered attention towards environment stimuli in horses in a poor welfare state (i.e. chronic back pain, depression-like state). This novel attention test could help to better identify factors inducing variation of attention, including intrinsic (e.g. welfare state, age…) but also environmental factors (e.g. riding practice), and ultimately to understand how to promote attention in horses.