Tracking the growth status of Chinese cabbage in real time is widely promising for exploring the effect of fertilization and increasing the yield of Chinese cabbage. In this paper, we constructed a quantitative computational model of Chinese cabbage growth status based on YOLOv5 network and multivariate linear fitting method using self-made Chinese cabbage growth stage dataset and leaf dataset. In order to accurately calculate the total fresh weight of Chinese cabbages, we used the YOLOv5 network to localize and identify the growth stages of individual cabbages, and to count the number of each of the four size types of leaves. Considering the leaf occlusion problem, here we use the leaf segmentation method based on HSV color space to calculate the image pixel ratio occupied by leaves. Together with the number of the four types of leaves as the independent variables, the equation for calculating the total fresh weight of Chinese cabbage was fitted by multivariate linear fitting. The fitting equation was tested by mAP and statistical manner, and the more accurate calculation effect was obtained.