In order to prevent the hazards of rock burst and the spontaneous fire in deep-level mine and ensure that the ventilation run safely and reliably, the variation trends and the distribution rules of the air flow in deep-level mine are analyzed. Based on the network flow theory and statistics principle, the density equations of the air quantities in the airways are established. With the maximum likelihood estimation (MLE), the rationality of the parameters in the air quantities distribution equations is verified. By means of the disjoint minimal path set algorithm, the reliability assessment models of mine ventilation network are founded. All the calculation processes are realized in the environment of MATLAB. The results indicate that the indexes of the quantitative analysis reflect the practical situation of the ventilation network in Daliuta coal mine.