Lameness assessment in Jeju horses presents unique challenges due to their distinct conformation and gait characteristics, and research in this area remains extremely limited. This study aimed to design an induced lameness model by comparing veterinary assessments with objective measurements obtained by a body-mounted inertial sensor system. A total of 63 Jeju horses (5.5 ± 2.3 years, 292.3 ± 22.3 kg) were included. Mild lameness was experimentally induced using temporary hoof foreign bodies. Gait asymmetry was independently evaluated by three equine veterinarians using the AAEP lameness scale and quantified with the Equinosis Lameness Locator. Primary parameters were Head Mean and Pelvis Mean displacements. Statistical analysis was conducted using Wilcoxon signed-rank test, and limb-specific detection were developed through machine learning approaches with cross-validation. Forelimb induction resulted in notable increase in asymmetry (Head Mean change: +3.73 ± 8.85 mm, p = 0.0201) whereas hindlimb induction did not yield significant changes (Head Mean change: +0.34 ± 5.93 mm, p = 0.9323). The sensor-based system achieved accuracies of 88.4% for forelimbs and 88.9% for hindlimbs. Veterinarians identified lameness in 98.2% of induced cases with overall agreement between veterinary assessments and sensor-based measurements reached 96.4%. Reference values were determined in sound horses (Head Mean 15.33 ± 7.19 mm and Pelvis Mean 5.72 ± 2.80 mm), showing substantial individual variation. This study proposes a practical and reproducible framework for objective lameness detection in Jeju horses. Results highlight that relative changes in gait symmetry, rather than absolute displacement values, provide the most diagnostic indicator. The model's simplicity and reliance demonstrate strong potential to complement veterinary decision-making and to improve future standardization of lameness evaluation in this native breed. Continued data collection is essential for further refining the thresholds and improving model precision.