Abstract: Objective To establish a rapid continuous monitoring method for qualitative determination of hepatitis B virus surface antigen ( HBsAg). Methods A rapid continuous detection program was established though exploring the conditions of reaction time for HBsAg test by amplified luminescent proximity homogeneous assay linked immunosorbent assay. The judgment rules for negative and positive samples were established by using the Classification and Regression Tree ( CART) based on 10 combinatorial classification variables. The judgment rules for suspicious results were established simultaneously according to the overlapping areas of negative and positive samples. The detectable limit of continuous monitoring method was verified, and the retest rates for negative results in continuous monitoring method and the first reading were compared. Results The incubation times for HBsAg detection of by amplified luminescent proximity homogeneous assay linked immunosorbent assay were determined to be 9 and 4 min in two tests ，respectively. The detections were performed once per minute for 3 consecutive times with 15 minutes of the final total reaction time. The judgment rule was established according to CART, by which the negative coincidence rate of sample verification was 100% and the positive coincidence rate was 97.5%. The risk indexes were 0.02 and 0.011 for training samples and validated samples respectively, both of which were less than 0.05. The detectable limit of the continuous reading protocol was verified to meet the requirements at the concentration of 0.1 IU/mL.The negative retest rate of the continuous monitoring method was 2.55%, which was significantly different from the retest rate of 7.64% in the first reading value protocol (X2 =4.215, P= 0.04). Conclusion The reaction time was shortened by continuous monitoring method, and the requirements of both qualitative accuracy and low retest rate were satisfied.