Novel object tests are often used to evaluate a horse's temperament by recording fear responses toward an unfamiliar object, but they might also be used as a part of welfare assessment. Various objects are used during these tests. This study aims to verify the use of different objects during a novel object test performed in the horse's stall. To this end, 54 horses and 4 objects (red-white striped cone [RWCONE]; a red-yellow plastic ball [BALL]; a black open umbrella [UMBR]; black-yellow striped cone [BYCONE]) were selected. To verify associations between behavior during the novel object test and undisturbed behavior at stable, baseline behavior profiling (4 × 10 minutes, spread over 4 days) was carried out. Thereafter, novel object tests were performed. Each object was presented for 10 minutes to the horses in their stalls, spread over 4 consecutive days. Each horse was exposed to the 4 objects in a semirandom order. The results reveal a higher frequency of object-related behavior in the presence of the umbrella (P = 0.0005), which might be caused by the color and the size of the object. No differences in object-related behaviors were found between the 2 cones which were colored differently. The age of the horse must be taken into account, as younger horses showed more pronounced reactions to the objects. When feed is available, horses showed less contact with the object, as they are not inclined to explore their environment. Sniffing behavior toward the object positively correlates with specific features of housing (large stall, visual and physical contact with other horses). A higher frequency of fear reactions to the objects is associated with more vocalizations during undisturbed behavior observations, which might be indicative of stress. The results show that associations between behavior during the novel object test and behavior at stable were present. However, the presence of roughage influences the horse's reaction and should be taken into account. In this case, the focus should be on the presence of fear reactions, such as rearing, not approaching the object or defecating, to correctly determine the animal's welfare state.