This work focuses on using EEG signals to detect mental commands through machine learning techniques. The goal is to aid in diagnosing psychiatric disorders and assisting individuals post-diagnosis. EEG data from 30 individuals were collected following a precise protocol. Volunteers perform tasks mentally, like moving a virtual box, allowing the AI model to learn corresponding EEG patterns. The dataset is cleansed and preprocessed for constructing, training, and testing the AI model. The project concludes with a Unity game connected to an AI model recognizing five mental commands. This approach shows potential for improving diagnostics and enabling daily activities through mental control.