This article explores machine learning (ML) methods to classify motor imagery (MI) tasks using electroencephalography (EEG) data. The process involves preprocessing the EEG data by applying a band-pass filter. Four hyperparameters were tested: the lower and upper frequency of the band-pass filter in the preprocessing stage and the minimum and maximum time of the epoch used in the classification stage after the MI trigger. The tests were conducted by adapting an MNE-Python example code for data classification with Common Spatial Patterns (CSP) and Linear Discriminant Analysis (LDA) methods using the Python MNE and scikit-learn libraries. The dataset used was the “EEG Motor Movement/Imagery Dataset”, which consists of 109 individuals with 64 electroencephalography channels and three runs on four different tasks. The tests involved evaluating how preprocessing hyperparameters affect classification accuracy. The results show that, for the classification of MI tasks, the lower frequency of the filter should be between 3 Hz and 13 Hz. In contrast, the upper frequency can be anything greater than 12 Hz. Moreover, at least a 0.2 s time frame is needed for classification using machine learning algorithms.