Beforehand and precise diagnosis of neurodegenerative diseases such as Alzheimer’s disease (AD) and frontotemporal dementia (FTD) is crucial for effective patient management and treatment planning. This study presents T-EEGNet, a novel deep learning model that integrates transformers with the EEGNet architecture to enhance the prediction of AD and FTD using electroencephalography (EEG) data. The model achieves a notable accuracy of 94 and 97%, respectively, in distinguishing AD and FTD from healthy controls (CN). By leveraging the temporal dependencies in EEG signals, the integration of transformers with EEGNet captures the subtle changes associated with neurodegenerative diseases, thus improving diagnostic accuracy and enabling earlier detection. The high accuracy of T-EEGNet highlights its potential as a reliable tool in clinical settings, aiding neurologists in the differential diagnosis of AD and FTD. This approach also contributes to the growing body of research on applying advanced machine learning techniques in medical diagnostics, showcasing the evolutionary impact of AI in healthcare. The successful implementation of T-EEGNet demonstrates the feasibility of using deep learning models for complex EEG signal analysis, paving the way for future innovations in the diagnosis and treatment of neurodegenerative diseases.