Objective: The research is to study the expression and distribution of matrix metalloproteinase (MMPs)-2 and -9 in the guinea pig cochlea after noise exposure, and to explore the role of MMPs in the blood-labyrinth-barrier (BLB). In addition, the role of MMPs inhibitor doxycycline in noise-induced BLB trauma was studied as well, which provides the basis for further studies and prophylaxis of noise-induced hearing loss. Methods: A total of 45 healthy adult guinea pigs were randomly divided into the control group (15 received intraperitoneal injection of 0.9% saline for 4 consecutive days), the noise-exposure group (15 exposed by 120 dB SPL white noise for 4 h per day for continuous 2 d, intraperitoneal injection of normal saline for 4 consecutive days) and the noise-exposure + doxycycline group (15 exposed by 120 dB SPL white noise exposure for 4 h per day for 2 consecutive days, and intraperitoneal injection of doxycycline 50 mg/kg/d for 4 consecutive days), respectively. Immunofluorescence staining, western blot, and real-time quantitative PCR were used to analyze the distribution and differential expression of MMP-2 and -9 in the stria vascularis of guinea pigs in comparison with the normal control group, noise only group, and noise & doxycycline treatment group. Immunofluorescence staining was used to observe the changes in tight junction (TJ) protein ZO-1 in stria vascularis in three groups and to investigate the effect of acoustic injury on TJs. And ABR tests were utilized to detect the hearing function of guinea pigs in the three groups. Intravenous Evans blue was administrated intravenously as an indicator of vascular leakage among three groups to study the changes in BLB permeability in context of acoustic injury. SPSS 22.0 was used for statistical analysis. Results: There was no significant difference in hearing function between the noise-exposure group and the noise & doxycycline group two hours after noise exposure. After seven, 14 and 28 days noise exposure, the hearing recovery of the noise & doxycycline treatment group was significantly greater than that of the noise-exposure group (P<0.05) . Immunofluorescence staining showed that there was only a small amount of MMP-2 and -9 in the stria vascular in the normal control group, and ZO-1 showed dense linear expression. While, in the noise-explore group, MMP-2 and -9 in the stria vascular was significantly elevated (P<0.05), and the configuration of ZO-1 became loose and discontinuous. However, the MMP-2 and -9 in the noise & doxycycline treatment group were not significantly different from the normal control group (P>0.05), which were significantly less than that in the noise-exposure group, and just a little break of ZO-1 was observed, however, the overall structure remained dense. The leakage of Evans blue from stria vascular capillary in the noise-exposure group was significantly increased, and the difference between the other two groups did not show any statistical significance (P>0.05). Conclusions: The damage of tight junction structure induced by MMP-2 and -9 may play an important role in BLB destruction. In addition, doxycycline can inhibit MMPs secretion, thereby, to some extent, protecting the integrity of BLB from acoustic injury, and contributing to the long-term hearing recovery.