Aflatoxin B1 (AFB1) is a predominant toxin detected abundantly and has been implicated in the etiology of human hepatocellular carcinoma. In this study， a novel quantitative recognition strategy of Aflatoxin B1 was proposed. The reaction medium for AFB1 was built using colloidal gold immuno-chromatography and information of the results was obtained using the technology of charge coupled device (CCD). An expressing method called color space was established to express the information of AFB1. The path can be used to obtain accurate measurement of the AFB1. The results showed that the color distance of T line and C line (AFB1 test card) was linearly correlated with the concentration (R2=0.986). The precision of the experiments was as follows: the relative standard deviation of the high concentration (20 μg/kg) was <2.7% for more than 97% dots， and that of the other two samples was <3.6% for more than 90% dots. This showed that the degree of coincidence of information in pixels between parallel color images(the same sample) was high. The maximum error between the linear results obtained by machine learning methods and the linear results obtained by 30 groups of parallel experiments is 5.93%， additionally， the results had no significant difference between RqM-AFB1 and HPLC test (P＞0.05). Thus， this method provides an effective way to improve fast quantitative and accurate identification of Aflatoxin B1.