Hyperspectral imaging technology was used to determine total flavonoid content distribution map on different color ginkgo biloba leaves (green, green yellowish, yellow). Firstly, 120 Ginkgo biloba leaves were used to collect hyperspectral image data cube and determine total flavonoids concentrations. Secondly, the average gray value of each band in hyperspectral image data cube was calculated as spectal data of Ginkgo biloba leaf. The spectal data and total flavonoids concentrations were used to bulid calibration model (R=0.9307) by using multiple stepwise regression method. Finaly, the calibration model was used to predict the total flavonoid content of each pixel in the hyperspectral image. This enables construction of a distribution map of total flavonoid content on the ginkgo leave. Distribution maps of total flavonoid content in three color groups (green, green yellowish, yellow) were calculated. Results indicate that the level of total flavonoid for green, green yellowish, and yellow were in an increasing order. Higher level of total flavonoids could be noticed at the margin area of the leaves, while lower levels occured at the area near the leafstalk of the ginkgo leaves. This research provided a method to determine major constituent of food and agricultural products.