Low-level viraemia (LLV) occurs in chronic hepatitis B (CHB) patients despite antiviral treatment, which may cause failed histological regression. Our study aimed to investigate the impact of different LLV types on fibrosis regression. The prospective study enrolled CHB patients with paired liver biopsies before and after 260 weeks of entecavir treatment. Fibrosis regression was defined by the Ishak score or P-I-R system. Patients were grouped as the SVR (HBV DNA < 20 IU/mL persistently) or LLV (HBV DNA between 20 and 2000 IU/mL), which were further grouped as very low-level viraemia (VLLV, HBV DNA < 50 IU/mL), occasionally LLV (OLLV, HBV DNA ≥ 50 IU/mL only once) and multiple LLV (MLLV, HBV DNA ≥ 50 IU/mL more than once). Logistic regression models were used to calculate the adjusted odds ratios (aORs) and 95% confidence intervals (CIs). The analysis included 111 CHB patients. In the SVR group (n = 54), 39 (72.2%) patients had fibrosis regression, which was higher than the LLV (56.1%, p = 0.080). The fibrosis regression rates for VLLV (30 patients), OLLV (17 patients) and MLLV (10 patients) were 70.0%, 52.9% and 30.0%, respectively. Compared with SVR, VLLV (aOR = 0.78; 95% CI: 0.28-2.21; p = 0.644) was not associated with fibrosis regression, but patients with non-VLLV (aOR = 0.27; 95% CI: 0.09-0.85; p = 0.025), especially with MLLV (aOR = 0.19; 95% CI: 0.04-0.97; p = 0.046) is significantly associated with hindered fibrosis regression. Our study suggests that patients with detectable serum HBV DNA levels higher than 50 IU/mL need to be monitored carefully, especially in those with more than once. Trial Registration: ClinicalTrials.gov identifiers NCT01938781 and NCT01938820.