This study identified the penetrability classification of the given compounds in biopharmaceutics classification system (BCS) through the establishment of mol. property based classification tree model.The Caco-2 cell penetration data collected from different literatures was used to constitute the training set.Classification tree model was built on the basis of the training set and applied in the classification test of the external test set, which was from the FDA BCS standard compoundsRules for the identification of BCS penetrability classification was: if the number of hydrogen bond donor was less than 4, the pos. Van der Waals polar surface area was less than 40.71 and the solvation energy was over -33.89, the compound had high penetrability, otherwise, low penetrability.The recognition correct rates of 105 compounds' training set and 17 compounds' external test set were 91.43% and 82.35%, resp.The mol. property classification modeling was successfully used in the FDA BCS standard compoundsThis study provided a convenient and feasible tool for the identification of medicinal penetrability of compounds