Objective To explore the effects of Xiangju capsule on aquaporin (AQP), bacterial biofilm formation and Toll-like receptor 4(TLR4)/ nuclear factor -κB(NF-κB) signaling pathway in rats with secretory otitis media. Methods A total of 60 rats were randomly divided into the normal group, the model group, the low-, medium- and high-dose groups of Xiangju capsule and positive control group, with 10 rats in each group. Except for the normal group,animal models of secretory otitis media were established in all other groups. During the modeling process, 2 rats died in the model group and 1 in the low-dose group of Xiangju capsule . After modeling, the low-, medium- and high-dose groups of Xiangju capsule were intragastrically administered with 0.36 g/100 mL, 0.54 g/100 mL and 0.72 g/100 mL of Xiangju capsule solution respectively for 14 consecutive days. The positive control group was intragastrically given 0.003 g/100 mL of prednisone, while the rest of the rats were intragastrically administered with the same volume of distilled water. The score of tympanic membrane inflammation, serum levels of AQP1, AQP4 and AQP5, bacterial biofilm, pathological morphology and protein levels of NF-κBp65 and TLR4 were compared among the groups. Results Compared with the normal group, the score of tympanic membrane inflammation, the protein levels of AQP4, NF-κBp65, TLR4 were increased in the model group, while the levels of AQP1 and AQP5 were decreased (P<0.05). Compared with model group, the score of tympanic membrane inflammation, the protein levels of AQP4, NF-κBp65 and TLR4 in low-, medium- and high-dose groups of Xiangju capsule were decreased, and the levels of AQP1 and AQP5 were increased (P<0.05). However, there was no difference between the medium-dose group of Xiangju capsule and the positive control group (P>0.05). Compared with the medium-dose group of Xiangju capsule, the tympanic membrane inflammation score, the protein levels of AQP4, NF-κBp65 and TLR4 were decreased in the high-dose group of Xiangju capsule, and the levels of AQP1 and AQP5 were increased (P<0.05). No bacterial biofilm formation was observed in the normal group. The bacterial biofilms in the model group presented cluster-like masses and were relatively large in area. In the low-dose group, fog-like bacterial biofilms could be observed. No bacterial biofilm aggregation could be observed in the medium-dose, high-dose groups of Xiangju capsule and the positive control group. In the normal group, the mucosal epithelial cells were neatly arranged, and blood vessels and fibrocytes were abundant, without obvious inflammation and hyperplasia. In the model group, the mucosal epithelial cells showed severe necrosis, the mucosae were thickened with a large amount of inflammatory cell infiltration, and the fibrous tissue hyperplasia was significant. In the low-dose group of Xiangju capsule, there was a small amount of inflammatory cell infiltration and a small amount of tissue hyperplasia. The histopathological tissues of the positive control group and the medium-dose group of Xiangju capsule were similar, with only a very small amount of fibrous tissue hyperplasia. The high-dose group of Xiangju capsule tended to be normal. Conclusions Xiangju capsule can inhibit the formation of bacterial biofilm and reduce AQP4, increase the levels of AQP1 and AQP5, and thereby control the condition of secretory otitis media. The mechanism may be related to inhibiting the activation of NF-κBp65 and TLR4 proteins.