Draw resistance and ventilation of cigarette are important indexes that affect the suction feeling, smoke and sensory quality of cigarette. In order to achieve the early prediction of draw resistance and ventilation of cigarette, the relevant factors were identified by analyzing the principle of draw resistance and ventilation, combined with a large number of data with different specifications and material characteristics. The prediction model of draw resistance and ventilation of cigarette was established by using multivariate adaptive regression splines (MARS) method. After verification, the normalized mean squared error (NMSE) of draw resistance prediction model was 0.276, and the average absolute error was 37.5 Pa; the NMSE of paper ventilation prediction model was 0.184, and the average absolute error was 0.91%; the NMSE of filter ventilation prediction model was 0.044, and the average absolute error was 1.27%. The results showed that the model could be applied to the prediction of cigarette draw resistance and ventilation in actual production, and provide a reference for cigarette materials matching in product design and production improvement.