Brain-Computer Interfaces (BCIs) are promising systems to allow direct communication between the human brain and external devices. Steady-State Visual Evoked Potentials (SSVEP) have been largely studied as BCI paradigm due to their high signal-to-noise ratio, compared to other BCIs. This study explores the application of machine learning methods to classify SSVEP signals, aiming to enhance the accuracy and efficacy of BCIs. The research conducts a comparison between Nearest Centroid, Decision Tree, Random Forest, Gradient Boosting, Support Vector Machine, and Multilayer Perceptron. The dataset comprises signals from an 8-channel wearable electroencephalogram (EEG) device, gathered from 40 healthy volunteers engaged in an SSVEP-based BCI task with 12 targets. Each subject’s session includes recordings using both wet and dry electrodes across 10 consecutive blocks. The assessment of machine learning techniques is based on mean accuracy, achieving 68.9%. Our findings provide insights into the effectiveness of these techniques for SSVEP pattern classification. Notably, all methods reach a similar performance. In part, the results are due to the fact that the signals are captured with wearable devices, a much more challenging scenario than conventional datasets.