The establishment of a fully functional and accurate pulmonary ventilation mechanism model is of great significance to simulate the real respiratory states of the specified human and configure respiratory therapeutic equipment (such as ventilators and simulated lungs). This paper proposed an alternative method to establish the digital twin mechanism model of pulmonary ventilation, which is a composition of the geometric and physical model mapping to the real respiratory process of humans. Through simulation, it could more intuitively reflect respiratory states and generate reliable respiratory data. The consistency of the digital twin of pulmonary ventilation was verified by the respiratory experiments. According to the verification results, the correlation coefficient and goodness of fit were 98.4% and 96.8%, respectively, in different respiratory patterns; 98.4% and 96.2%, respectively, in different examinees. It can be concluded that the digital twin of pulmonary ventilation could well mimic various respiratory patterns of different human bodies. Then it can generate reliable data to support the parameter settings of the respiratory therapeutic devices and auxiliary respiratory devices such as simulated lungs and ventilators.