Abstract: In response to the characteristics of the measured data on the roadway ventilation frictional resistance coefficient, a Principle Component Analysis (PCA)-BP neural network prediction model is established. The PCA method is used to reduce the dimensionality for seven factors that affecting the roadway ventilation frictional resistance coefficient, such as support type, section shape, roadway width, roadway height, peripheral length of the supported part, roadway cross-sectional area, and roadway length. The contribution rates of the factors after dimensionality reduction are sorted and screened to obtain three principle component indicators (F₁, F₂ and F₃), which are used as neurons in the input layer of the BP neural network. The PCA-BP neural network model is trained and tested using measured data, and the test results are compared with the test resutls of the Support Vector Machine Regression (SVM) model and the BP neural network model. The results show that the average accuracies of the full-factor BP neural network prediction model and the SVM prediction model are 92.942 0% and 93.023 5%, respectively, while the average accuracy of the PCA-BP prediction model reaches 96.432 5%. The PCA-BP neural network model not only simplifies the structure of the network, but also improves it's generalization ability, which resulting in smaller prediction error and higher accuracy, and provides an effective method for obtaining the ventilation frictional resistance coefficient of the roadway more accurately.