Attention-Deficit/Hyperactivity Disorder (ADHD) often impairs academic performance, and while tools like the Continuous Performance Test II (CPT-II) assess attention, few systems offer immediate support to maintain focus during learning lessons. This study introduces a closed-loop Brain-Computer Interface (BCI) that delivers personalized and real-time haptic feedback to enhance sustained and selective attention in students during lectures. Using Enophone’s 4-channel Electroencephalography (EEG) and an Arduino-based haptic module, the BCI system monitors brain activity and delivers subtle vibrations when Machine Learning (ML)-predicted attention lapses. To train the ML EEG-based, 20 participants completed an experimental protocol consisting of a 1-minute Eyes Open (EO) baseline followed by a 9-minute CPT-II test; an Attention Score (AS) was calculated considering the inverse of the reaction times fitting a Power Spectral Density (PSD)-based Multiple Linear Regression (MLR) model with a 0.72 R2. Later, the non-invasive Bluetooth-based neurofeedback system was tested with 26 additional participants watching a 12-minute educational video, divided into three groups: (1) Control, (2) MLR-based neurofeedback, (3) Engagement Index-based neurofeedback. While task performance and self-reported attention did not differ significantly between groups, EEG-based engagement (p < 0.001), fatigue (p < 0.01), and excitement (p < 0.05) indices were significantly higher in the neurofeedback groups when compared to the control group. Although students with clinically diagnosed ADHD were not included in this study, results from neurotypical participants validate the system’s real-time functionality and attention-enhancing potential. This BCI system represents a promising step towards adaptive tools that enhance attention and improve learning experiences.