Abstract: In view of problems of poor practicability, poor stability and low accuracy of existing monitoring methods for underground personnel violation entry, a passive monitoring method for underground personnel violation entry based on channel state information of WiFi network was proposed. The method includes training phase and testing phase. In the training phase, channel state information data under conditions of somebody entry and nobody entry is collected respectively, and the collected data is preprocessed through outlier elimination and filtering. Then the preprocessed data is constructed into eigenvalue to establish discrimination model. In the testing phase, the collected data is preprocessed to construct eigenvalue, which is input into the discrimination model established in the training phase, so as to realize judgment of personnel violation entry. The experimental results show that accuracy of the method is 99.31%.