Objective To describe perioperative antimicrobial use in horses undergoing elective arthroscopy. Study design. Retrospective study. Sample population. Horses that underwent elective arthroscopy at one institution between July 2016 and May 2018, excluding those with a suspected infectious orthopedic disease or with a comorbidity that may have impacted prophylactic antimicrobial use decisions. Methods Medical records were reviewed to evaluate preoperative, intraoperative, and postoperative antimicrobial drug selection, dose, and timing. Associations between body weight and underdosing were evaluated by using analysis of variance, χ2 test was used for categorical comparisons, and least squares fit was used to evaluate factors associated with duration of postoperative antimicrobials. Results Among 150 horses, 149 (99.3%) received systemic preoperative antimicrobials. Only 53 (40.2%) horses were administered doses within 60 minutes of surgical incision. First incision was performed more than two half‐lives after administration of sodium penicillin in 46 of 131 (35.1%) horses but in only 1 of 106 (0.8%) horses that received trimethoprim‐sulfadoxine. Body weight was associated with underdosing for penicillin (P = .0075) and trimethoprim‐sulfadoxine (P = .002) but not gentamicin (P = .92). Twenty‐six (17%) horses received one postoperative antimicrobial dose, while antimicrobials were continued in hospital for a mean of 22.3 ± 4.4 hours after surgery in the other 123 horses. Among the 149 discharged horses, 115 (77.2%) were prescribed antimicrobials after discharge (range, 3‐10 days; median, 3 days, interquartile range, 0 days). Conclusion Deviations from common recommendations were apparent and provide evidence for the requirement to develop interventions to optimize perioperative prophylaxis. Clinical significance Perioperative antimicrobial use practices should be regularly assessed to provide a benchmark and identify areas for intervention.