Objective: To compare the clinical efficacy of a recombinant bovine basic fibroblast growth factor (rb-bFGF) gel and a gel matrix in the treatment of moderate dry eye. Methods: It was a prospective random double-blind controlled study. One hundred patients diagnosed as moderate dry eye in Eye Institute and Affiliated Xiamen Eye Center of Xiamen University, Union Hospital Affiliated to Tongji Medical College of Huazhong University of Science and Technology, Beijing Tongren Hospital, Capital Medical University, Eye & ENT Hospital of Fudan University and Zhongshan Ophthalmic Center from August 2015 to April 2019 were divided into two groups: experimental group and control group. Two groups of patients were allocated to receive either a rb-bFGF gel or a gel matrix 4 times per day for 4 weeks. Subjective symptoms, break-up time of the tear film (BUT), Schirmer Ⅰ test (SⅠt) and corneal fluorescein sodium staining were assessed at baseline, 2 and 4 weeks after treatment. Bulbar impression cytology was evaluated at baseline and 4 weeks after treatment. Irritation of the rb-bFGF gel and the gel matrix was estimated after treatment. T test, Wilcoxon signed-rank test or Mann-Whitney U test was used for quantitative data, and Chi-square test was used for enumerative data. Results: Eighty-four subjects were included for statistical analyses after the exclusion of 16 subjects who were lost for followup, with an age of 43±14 years. There were 42 cases in the experimental group and the control group, respectively. There was no statistically significant difference between the two groups in demographic baseline characteristics before treatment ( P >0.05). The total score of subjective symptoms was 7.17±3.60 and 5.95±3.25 at 2 and 4 weeks after therapy in the experimental group, which were lower than 9.48±3.88 before treatment ( t =6.226, 6.563; both P <0.05); in the control group, it was 7.01±3.25 and 6.32±3.85 at 2 and 4 weeks after treatment, with a significant reduction in comparison with that before treatment (9.15±3.58; t= 4.693, 4.726; both P <0.05). The median (lower quartile, upper quartile) BUT was 4.00 (2.40, 5.00) s and 4.64 (3.00, 5.00) s at 2 and 4 weeks after therapy in the experimental group, which were longer than 3.72 (2.00, 4.39) s before treatment ( Z =-2.485, -3.152; both P <0.05). The BUT was 4.41 (2.79, 5.12) s at 2 weeks after therapy in the control group, which was of no statistical difference compared with 3.89 (2.09, 4.25) s before treatment ( Z =-1.953, P >0.05). The BUT was 5.21 (3.00, 5.02) s at 4 weeks after therapy in the control group, which was longer than that before treatment ( Z =-2.485, P <0.05). The SⅠt score was 7.31 (3.75, 10.00) mm and 8.50 (4.00, 11.00) mm at 2 and 4 weeks after therapy in the experimental group, which were significantly higher than 6.69 (2.00, 8.13) mm before treatment ( Z =-2.031, -2.236; both P <0.05); in the control group, it was 6.82 (2.00, 8.25) mm and 6.86 (3.00, 9.25) mm at 2 and 4 weeks after therapy, which were not significantly increased compared with 6.50 (2.00, 7.75) mm before treatment ( Z =-0.179, -1.161; both P >0.05). The corneal fluorescein sodium staining points were 5.00 (2.00, 5.00) and 3.71 (0.00, 5.00) at 2 and 4 weeks after therapy in the experimental group, which were significantly lower than 7.10 (5.00, 7.00) before treatment ( t =-2.895, -4.639; both P <0.05); those in the control group were 5.52 (0.00, 7.00) and 6.19 (0.75, 6.25) at 2 and 4 weeks after treatment, with a significant reduction in comparison with 8.90 (5.00, 10.50) before treatment ( t= -2.776, -1.991; both P <0.05). The differences in the average total score of subjective symptoms, BUT, SIt, and corneal fluorescein sodium staining points between both groups were not statistically significant at each time point. The impression cytology grade was decreased from 1.72 (1.00, 2.00) before treatment to 0.94 (0.00, 2.00) at 4 weeks after therapy in the experimental group ( Z =-2.803, P <0.05). The staining grade of conjunctival imprinted cells in the control group was 1.42 (1.00, 2.00) at 4 weeks, which showed no statistical significance compared with 1.56 (1.00, 2.00) before treatment ( Z =1.195, P >0.05). The impression cytology grade was significantly reduced in the experimental group compared with the control group at 4 weeks after treatment ( Z =-3.308, P <0.05). The number of goblet cells was 10.90 (5.00, 20.00) at 4 weeks after therapy in the experimental group, which was significantly higher than 6.30 (5.00, 8.00) before treatment ( Z =-2.383, P <0.05); in the control group, it was 8.36 (4.00, 12.00) at 4 weeks after treatment, with no significant increase in comparison with that before treatment [7.55 (5.00, 11.00)] ( Z =-0.095, P >0.05). The number of goblet cells was not significantly increased in the experimental group compared with the control group at 4 weeks after treatment ( Z =-1.162, P >0.05). Most patients indicated that the drug was non-irritating, and no patient had intolerable irritation affecting daily lives at 4 weeks after therapy; there was no difference between the two groups ( Z =-0.290, P >0.05). Conclusions: Both the rb-bFGF gel and the gel matrix can effectively improve the symptoms and signs of moderate dry eye. However, compared with the gel matrix, the rb-bFGF gel shows obvious advantages in promoting conjunctival epithelial cell repair and increasing the number of goblet cells. (Chin J Ophthalmol, 2021, 57: 930-938) .