People with motor disabilities will be facing difficulties to interact with the environment with the support of their (PNS) peripheral nervous system. Brain–computer interface (BCI) systems use activity of the brain's cortex to control prosthetic devices like robotic arm which would be highly helpful in passing information to them EEG can be utilized to detect activations of the human brain indicating the motor activity and it is an upcoming idea in the field of BCI. This idea aims at developing a system that records the motor execution signals during the right- and left-arm & hand movements from 7 channels placed on frontal lobes. The alpha and beta signals are of interest because they correspond to the voluntary motor movements. Our work aims at acquiring EEG signals, pre-processing to denoise and only the signals from the channels of interest were extracted. Along with features like mean, minimum, maximum, two other features - log energy entropy and Katz fractal dimension are investigated and their effects on the classifiers SVM and Random Forest were studied. The classifiers' performance is analysed when all 7 channels out of 21 are considered. Classification of EEG signals corresponding to motor execution tasks gave accuracy value of (93 ± 2%)% along with F1- score of 0.98 for Random Forest and accuracy (89±5)% and F1- score of 0.89 for SVM over 7 channels of frontal lobe EEG, Proving that Random forest classifier can be utilized for classification of EEG signals.