Hepatotoxicity, drug-induced liver injury, and competitive Cytochrome P-450 (CYP) isozyme binding are serious problems associated with drug use. It would be favorable to avoid or to understand potential CYP inhibition at the developmental stages. However, current in silico CYP prediction models or available public prediction servers can provide only yes/no classification results for just one or a few CYP enzymes. In this study, we utilized a rule-based C5.0 algorithm with different descriptors, including PaDEL, Mold(2), and PubChem fingerprints, to construct rule-based inhibition prediction models for five major CYP enzymes-CYP1A2, CYP2C9, CYP2C19, CYP2D6 and CYP3A4-that account for 90% of drug oxidation or hydrolysis. We also developed a rational sampling algorithm for the selection of compounds in the training data set, to enhance the performance of these CYP prediction models. The optimized models include several improved features. First, the final models significantly outperformed all of the currently available models. Second, the final models can also be used for rapid virtual screening of a large set of compounds due to their ruleset-based nature. Moreover, such rule-based prediction models can provide rulesets for structural features related to the five major CYP enzymes. The five most significant rules for CYP inhibition were identified for each CYP enzymes and discussed. An example was chosen for each of the five CYP enzymes to demonstrate how rule-based models can be used to gain insights into structural features that correspond with CYP inhibitions. A newer version of the freely accessible CYP prediction server, CypRules, is presented here as a result of the aforementioned improvements.