Objective To carry out the clinical validation of a domestic HBsAg kit to evaluate its application value. Methods 543 serum samples with HBsAg ELISA values of S/CO ≥ 0. 7 were tested by HBsAg confirmatory test. Specific anti-HBs reagent and control reagent were added separately into duplicate wells of HBsAg ELISA plate, in which test sample was also added. After incubation at 37℃, HBsAg was detected by routine ELISA, and the inhibition rate was calculated using absorbanee (A) result of anti-HBs reagent well and control reagent well according to the provided formula. The sample was confirmed as HBsAg positive when the inhibition rate was≥50%. For HBsAg weakly positive samples, "prolonged confirmatory test" (conjugate reaction time was prolonged to 120 rain) was applied to increase the sensitivity. 39 samples were randomized selected for testing and comparison with Abbott Murex confirmatory test. Results 543 serum samples in total were tested by the confirmatory test. Among the 504 cases which showed positive reaction in screening HBsAg ELISA, 89 ( 17. 7% ) were confirmed as negative. According to their S/CO value of the screening HBsAg test, the ratio of negative cases / tested eases in the group were:S/CO≤<5.0, 87/143 (60. 8% ) ;5.0 < S/CO ≤ 10. 0,0/25 (0) ;10. 0 < S/CO ≤ 15.0, 1/21 (4. 8% ) ;15.0 < S/CO ≤ 20. 0, 1/23 (4. 4% ) ;S/CO 20. 0, 0/292(0). Among 39 cases with negative HBsAg (0. ≤<S/CO≤1.0), 2 cases were confirmed as positive by prolonged confirmatory test. 39 cases with positive HBsAg ( 1.0 < S/CO≤20. 0) were confirmed as positive both by domestic and Murex confirmatory tests. Conclusions The domestic confirmatory test kit is suitable for clinical use to confirm the results of samples tested by domestic HBsAg ELISA kits. Weakly positive (S/CO≤5.0) serum samples detected by primary HBsAg ELISA should be tested with confirmatory test, which is not necessary for strongly positive (S/CO 20. 0) samples. More clinical studies should be performed to establish a proper cut-off for the S/CO value of weakly positive samples.