Abstract: The abnormal change of ventilation system air volume is also known as the resistance variant fault,and is caused by roadway falling deformation,air-door opening and closing or broken,fan performance declining,roadway ex- tension or scrap,tramcar running,cage lifting and coal silo emptying et al. The network topology location of ventilation variant fault and its equivalent wind resistance are determined according to the numerical changes of air volume sen- sor. The above-mentioned method holds great influence on intelligent mine ventilation management and security capa- bilities improvement. Ventilation network possesses better self-adaptability and robustness,the mine ventilation simula- tion system MVSS is used to generate the training samples. SVM-based classifiers and regressors are constructed re- spectively to diagnose the fault location and the equivalent wind resistance. Ventilation system SVM classification and regression,which is only characterized by air volume,is an ill-posed problem. This study shows that the accuracy of the diagnosis is related to the number of sensors and the degree of dispersion of the location,not to the network topology sensitivity of the roadway where the sensor is located and the roadway air volume.