In a rodent model of acute aminoglycoside ototoxicity to unilateral inner ear, physical abnormalities, such as nystagmus and postural alteration, are relieved within a few days by neural compensation. However, how exploratory behavior changes after unilateral ototoxic damage in home-cage environment in mice have not yet been elucidated. In this study, exploratory behavior after kanamycin-induced unilateral inner ear injury was examined in a cage with wood-shavings as natural bedding. To this end, we developed machine learning environment with robust detection model for evaluation of drug efficacy and toxicity. In sequential images from video recording, our deep learning model detected the mouse as an object with 99.9% accuracy. After evaluation of the detection accuracy, tracing of the mouse movement revealed that total distance moved in 15 minutes was increased 3 days after surgery. Moreover, the injured mouse turned frequently toward healthy side upto 17 days after surgery. Our mouse model of unilateral inner ear dysfunction and its analysis strategy is useful to evaluate neuronal compensatory process and screening of drug with therapeutic potency.