Through detecting the standard preparation with series of concentration to indirectly calculate the anti-HBs concentration of the serum samples, a suitable anti-HBs quantitative method for our laboratory was found after comparing the two kinds of methods.Detecting the anti-HBs standard preparation with series of concentration by RIA method, standard curvilinear equations were obtained by the means of fitting the detected result and the corresponding concentration by log-log model and exponential curve model respectively. Then the fitting efficiency of two curves was compared. By calculating the concentrations of the reference using two standard curvilinear equations, we can compare the accuracy of two quantitative methods.The error mean square of the exponential curve model is low as 1.2971 and the determinate coefficient is close to 1 with the value of 0.9904. The average concentrations (n=6) of the detected reference calculated by two curvilinear equation with the actual concentration of 30.0 mIU/mL are (32.28 +/- 1.06) and (31.91 +/- 1.06) mIU/ mL respectively. The concentration calculated by exponential curve model is only 6.37% higher than the actual concentration.Fitting by exponential curve model is more practical to estimate the actual concentration of the serum samples those will be detected. It can be used as an optimal quantitative method to detect anti-HBs concentration.