Objective The aim of this study was to evaluate the effectiveness of high-dose corticosteroid (120 mg prednisolone equivalent daily) in Bell’s palsy compared with low-dose corticosteroid (60 mg PSL equivalent). Methods A single-center retrospective observational study was performed. We included adult Bell’s palsy patients who were treated within 7 days after disease onset. We compared high- and low-dose corticosteroid for the non-recovery rate at 6 months after disease onset using inverse probability-weighted propensity score analysis (IPW-PS). Results A total of 368 Bell’s palsy patients (281 in the high-dose and 87 in the low-dose group) were included. The non-recovery rate without IPW-PS was 13.8% in the low-dose and 8.2% in the high-dose group. After IPW-PS adjustment, the non-recovery rate was 13.1% in the low-dose and 7.8% in the high-dose group (difference = −5.28%, 95% confidence interval [CI] −12.7% to −2.1%, p = 0.040). High-dose corticosteroid decreased the non-recovery rate in severe Bell’s palsy patients with a Yanagihara score of 0–10 (difference = −16.1%, 95% CI −38.5% to −6.2%, p = 0.012), but did not decrease in moderate Bell’s palsy patients with a Yanagihara score of 12–18 (difference = −2.0%, 95% CI −11.0% to 7.0%, p = 0.591). Subgroup analysis revealed that the efficacy of high-dose corticosteroids was higher when patients were treated within 3 days after disease onset, but not when patients were treated at 4 days or later after disease onset. Conclusions Physicians would be better to treat severe Bell’s palsy patients with high-dose corticosteroids when the patients are treated within 3 days after disease onset.