Acanthopanax senticosus is widely cultivated in China, Japan, South Korea, and Russia, and is extensively researched worldwide. Illegal activities, particularly adulteration have been fostered due to its large market demand and high price. This study successfully identifies authentic and adulterated Acanthopanax senticosus samples using a portable mass spectrometer combined with chemometrics. Different data processing and feature selection methods were evaluated for authenticity and adulteration models, with SNV-VIP being selected as the optimal method for both models. Excellent classification performance was achieved using Orthogonal Partial Least Squares Discriminant Analysis (OPLS-DA), K-nearest Neighbors (KNN), Random Forest (RF), and Convolutional Neural Networks-Long Short-Term Memory-Attention (CNN-LSTM-Attention) models, with prediction accuracies for the authenticity model validation set being 1.00, 1.00, 1.00, and 1.00, respectively. The prediction accuracies for the adulteration model validation set were 0.86, 0.73, 0.89, and 0.92, respectively. Additionally, the models identified the most important ions based on variable importance. The combination of portable mass spectra (PMS) and chemometrics provides a simple, rapid, and reliable method for identifying adulteration in Acanthopanax senticosus samples. This study provides a new basis into the application of PMS for adulteration detection in the food field.