It is currently unclear what impact serum microRNA-122 (miR-122) levels have on clearance of hepatitis B virus (HBV) surface antigen (HBsAg) in HBV-infected patients who had not received antiviral therapy. The current study evaluated the impact of serum miR-122 levels on HBsAg seroclearance in 367 consecutive HBV-infected patients who had not received antiviral therapy between their initial and last visit, and investigated the predictive factors of HBsAg seroclearance. Cumulative HBsAg seroclearance rates were 13.5%, 32.0%, and 37.4% after 10, 20, and 30 years, respectively. The yearly incidence of HBsAg seroclearance over the investigated 30-year period was 1.25%. A significant and strong correlation was observed between serum miR-122 and HBsAg levels. Moreover, there was a significant correlation between serum miR-122 levels and the levels of HBV DNA, hepatitis B e-antigen, and HBV core-related antigen. The HBsAg seroclearance rate in patients with a <1.0-fold change of serum miR-122 levels was significantly higher than in those with a ≥1.0-fold change. Multivariate analysis identified age (≥30 years), HBV DNA levels (<2.2 log U/mL), HBV genotype (non-C), and serum miR-122 levels (<1.0-fold change) as significant predictors of HBsAg seroclearance. Our results indicated that serum miR-122 level is an important predictor of HBsAg seroclearance in Japanese patients who do not receive antiviral therapy. Understanding the complexity of the interactions among various virus-related and host-related factors could potentially help in the design of new therapies that enhance HBsAg seroclearance.