Only little is known about behaviour and stress responses in horses with gastric ulceration, despite the high prevalence of this condition. Our objectives in the present study was to (i) describe the severity of gastric ulceration in horses, housed under relatively standardised conditions, and (ii) to investigate whether horses with severe glandular gastric ulceration have increased baseline and response concentration of stress hormones and behave differently than control horses. We investigated stomachs of 96 horses at one stud, and compared an ulcer group (n = 30; with severe lesions in the glandular mucosa) to paired controls (n = 30; free from gastric ulcers). Baseline and response concentrations of faecal cortisol metabolites (FCM), heart rate and behaviour were measured in a novel object test (NOT, Day 1) and behaviour during postponed feeding (PF, Day 2). Glandular lesions occurred in 55.2% and non-glandular lesions in 40.6% of the horses. The amount of starch in the feed (P = 0.006) and paternal stallion (P = 0.031) influenced ulceration in the non-glandular region only; it should be noted that our study does not allow for separating hereditary from environmental influences, as offspring may be e.g. trained differently dependent on breeding line. Ulcer horses pawed more (P < 0.001) and ate quicker (P = 0.050) during PF. Although displayed by ulcers horses only during PF, we failed to demonstrate a significant association between glandular gastric ulceration and crib-biting/weaving; the total number of horses with these types of abnormal behaviour was low (n = 5). Behaviour and heart rate did not differ between groups in the NOT. Baseline concentration of FCM was similar (P = 0.79), however, ulcer horses responded stronger to novelty than controls (26% higher FCM; P = 0.018). We conclude that the prevalence of gastric ulcers was high, and our results suggest different factors affecting ulceration in the glandular versus the nonglandular region of the horse stomach. Obvious external signs (e.g. poor body condition) identifying ulcer horses were absent. Horses with severe glandular ulcers had a higher stress hormone response to novelty, thus they were more stress sensitive. Consequently, management evoking stress in horses should be reduced to dampen the development of glandular ulceration, or to protect horses with this condition.