This paper introduces an integrated software system that uses a machine learning model to classify sour and sweet tastes based on electroencephalography (EEG) signals. The system utilizes PyQt5 to provide an intuitive graphical user interface (GUI), enabling users to log in, upload, store, and reuse EEG data from various formats, including CSV, MAT, EDF, and TXT. The system supports key signal preprocessing steps, including noise filtering, rate limiting, detrending, and baseline filtering, and visualizes both raw and processed signals. A notable feature is the ability to extract features using wavelet analysis (db4, db6) and calculate spectral power in the standard EEG frequency bands (Alpha, Beta, Theta, Delta, Gamma), which are displayed via corresponding charts. Crucially, the system enables the training and evaluation of multiple classification models, including SVM, KNN, ANN/MLP, and RF, on a labeled EEG dataset of “sweet” and “sour” tastes, clearly displaying the prediction results. Trained models can be saved in the .pkl format for easy redeployment. All user data and EEG files are securely stored in a MySQL database, ensuring personalization and security. Experimental results confirm the system’s efficiency in processing EEG data, making signal processing and machine learning in taste classification accessible to non-experts. The system has the potential to be expanded for other brain-computer interface (BCI) applications, such as emotion recognition or attention state detection.