The invention relates to the field of health risk assessment testing strategies, in particular to a method for predicting the skin permeability coefficient of organic chemicals. On the basis of obtaining the molecular structure of the compound, the quantitative structure-activity relationship (QSAR) is used to construct a prediction model by calculating the descriptors that characterize the structural features. Compared with the traditional test method for measuring skin penetration parameters, it is in line with animal welfare protection and reduces test time and cost. , which can quickly and effectively predict the skin permeability coefficient. The present invention is strictly in accordance with the 5 standards proposed by the Organization for Economic Co-operation and Development (OECD) for the construction and use of QSAR models, by calculating the physical and chemical properties, electrical properties, topology and quantum chemical parameters of compounds as predictive descriptors, and using K-S grouping to The original data is classified, and 7 optimal descriptors are screened out. Using the clear, simple, fast, and transparent GA-MLR algorithm, the model application domain is clear, and it has good fitting effect, robustness, and predictive ability. The skin permeability coefficient prediction model can accurately and efficiently predict the skin permeability coefficient of compounds, and provides an effective method for the health hazard assessment of organic compounds.