This work proposes an assistive system for everyday activities composed by a brain machine interface (BMI) based on P300 to choose a predefined task, a robot arm to perform the chosen task, and a stereo vision subsystem developed with two cameras for object recognition and coordinates calculation. The system was tested with eight healthy subjects; its results were greater BMI accuracies, lower 3D coordinates calculation error, and lower task execution time than similar systems. However, it should be tested with disabled subjects to provide more reliable end-user results. Regardless, this system is suitable to assist healthy subjects for performing reaching task to grasp objects in daily activities, and the intuitive interface would be useful for disabled subjects. © 2019 IEEE.