On the basis of oxidative decoloration of bromopyrogallol red(BPR) with H_ 2O_ 2, catalyzed by horseradish peroxidase(HRP), and the sequential injection renewable surface technique(SI-RST), a highly sensitive optical-fiber sensor spectrophotometric method for the enzymatic determination of hydrogen peroxide was proposed. By coupling with a glucose oxidase(GOD)-catalyzed reaction, the method was used to determine glucose in human serum. The considerations in system and flow cell design, and factors that influence the determination performance are discussed. With 100 μL of sample loaded and 0.6 mg of bead trapped, the linear response range from 5.0×10 -8 to 5.2×10 -6 mol/L BPR with a detection limit(3σ) of 2.5×10 -8 mol /L BPR, and a precision of 1.1% RSD(n=11) and a throughput of a 80 samples per hour can be achieved. Under the conditions of a 8.7×10 -6 mol/L BPR substrate, 0.04 unit/mL HRP, 600 s reaction time and a reaction temperature of 37 ℃, the linear response range for H_ 2O_ 2 was from 5.0×10 -8 to 7.0×10 -6 mol/L with a detection limit(3σ) of 1.0×10 -8 mol/L and a precision of 3.7% RSD(n=11). The linear response range by coupling with a GOD-catalyzed reaction was from 1.0×10 -7 to 1.0×10 -5 mol/L. The method was directly applied to determine glucose in human serum. Glucose contents obtained by the proposed procedure were compared with those obtained by using the phenol-4-AAP method, the error was found to be less than 3%. On the basis of oxidative decoloration of bromopyrogallol red (BPR) with H 2 O 2, catalyzed by horseradish peroxidase (HRP), and the sequential injection renewable surface technique (SI-RST), a highly sensitive optical-fiber sensor spectrophotometric method for the enzymatic Determination of hydrogen peroxide was proposed. By coupling with a glucose oxidase (GOD) -catalyzed reaction, the method was used to determine glucose in human serum. The considerations in system and flow cell design, and factors that influence the determination performance are discussed. With 100 μL of sample loaded and 0.6 mg of bead trapped, the linear response range from 5.0 × 10 -8 to 5.2 × 10 -6 mol / L BPR with a detection limit (3σ) of 2.5 × 10 -8 mol / L BPR , and a precision of 1.1% RSD (n = 11) and a throughput of 80 samples per hour. Under the conditions of a 8.7 × 10 -6 mol / L BPR substrate, 0.04 unit / mL HRP, reaction time and a reaction temperature of 37 ° C, the linear response range for H_ 2O_ 2 was from 5.0 × 10 -8 to 7.0 × 10 -6 mol / L with a detection limit (3σ) of 1.0 × 10 -8 mol / L and a precision of 3.7% RSD (n = 11). The linear response range by coupling with a GOD-catalyzed reaction was from 1.0 × 10 -7 to 1.0 × 10 -5 mol / L. The method was directly applied to determine glucose in human serum. by using the phenol-4-AAP method, the error was found to be less than 3%.