Twenty stock-type horses (589 ± 126 kg BW; 13 ± 8 yr) were used in a completely randomized design for 28-d to evaluate the impact of a joint supplement on gait kinematics, inflammation, and cartilage metabolism. Horses were stratified by age, sex, body weight (BW), and initial lameness scores and were randomly assigned to one of two dietary treatments consisting of either a 100-g placebo top-dressed daily to 0.6% BW (as-fed) commercial concentrate (CON; = 10; SafeChoice Original, Cargill, Inc.), or an oral joint supplement (SmartPak Equine LLC) containing glucosamine, chondroitin sulfate, hyaluronic acid, methylsulfonylmethane, turmeric, resveratrol, collagen, silica, and boron (TRT; = 10). Horses were group-housed with ad libitum access to coastal bermudagrass hay () and allowed to graze pasture 2 h/d. Horses were exercised progressively 4 d/wk at 45 min each. On days 13 and 27, blood was harvested followed by a 19.3-km exercise stressor on concrete. Horses traveled at the walk, with no more than 15 min at the trot. Every 14 d, BW and BCS were recorded, and blood was collected for plasma prostaglandin E (PGE), serum collagenase cleavage neopeptide (C2C), carboxypropeptide of type II collagen (CPII), and chondroitin sulfate 846 epitope (CS846) analysis. Kinematic gait analysis was performed every 14 d (Kinovea v.0.8.15) to determine stride length (SL) and range of motion (ROM) of the knee and hock at the walk and trot. Data were analyzed using PROC MIXED of SAS. All horses increased BW and BCS over time ( ≤ 0.01). Hock ROM increased in TRT horses ( ≤ 0.02) at the walk and tended to increase at the trot compared to CON ( = 0.09). At the walk, SL and knee ROM increased over time, independent of dietary treatment ( ≤ 0.01); no time effect was observed at the trot ( > 0.15). Regardless of treatment, C2C and CPII increased over time ( ≤ 0.05) and no effect was observed for CS846 or PGE ( > 0.12). In response to the exercise stressor, CPII and PGE decreased ( ≤ 0.05) from day 13 to 14, and CS846 and PGE tended to decrease ( 0.10) from day 27 to 28, independent of dietary treatment. In conclusion, hock ROM at the walk and trot was most sensitive to dietary treatment. Supplementation did not alter biomarker concentration of collagen metabolites or systemic inflammation in the 28-d period, but a future study utilizing arthrocentesis may be warranted to specifically evaluate intra-articular response to dietary treatment.