This study compares clinical characteristics between induction with thiopental/guaifenesin and propofol/ketamine in Thoroughbred racehorses anesthetized with sevoflurane and medetomidine. Clinical records of 214 horses that underwent arthroscopic surgery between 2015 and 2016 were retrospectively retrieved. Horses were premedicated with medetomidine and midazolam to sedate at the adequate level for smooth induction, and then induced with either thiopental (4.0 mg/kg) and guaifenesin (100 mg/kg) in Group TG (n=91) or propofol (1.0 mg/kg) and ketamine (1.0 mg/kg) in Group PK (n=123). Anesthesia was maintained using sevoflurane with constant rate infusion of medetomidine. Quality of induction/recovery, sevoflurane requirement, cardiovascular function and recovery characteristics were evaluated. Anesthetic induction scores (median, range) for Group TG (5, 2–5) and Group PK (5, 2–5) were not significantly different. There were no significant differences in end-tidal sevoflurane concentration (mean ± standard deviation) between Group TG and Group PK (both 2.4 ± 0.2%). Dobutamine infusion rate (µg/kg/min) required for keeping mean arterial blood pressure (MAP) above 70 mmHg in Group PK (0.43, 0.10–1.40) was significantly lower than in Group TG (0.67, 0.08–1.56). Recovery score in Group PK (5, 2–5) was significantly higher than in Group TG (4, 2–5). Both propofol/ketamine and thiopental/guaifenesin provided a smooth induction of anesthesia. Moreover, induction with propofol/ketamine resulted in lower dobutamine requirements for keeping MAP above 70 mmHg during maintenance, and better quality of recovery. Induction with propofol/ketamine would be preferable to thiopental/guaifenesin in Thoroughbred racehorses anesthetized with sevoflurane and medetomidine during arthroscopic surgery.