In the master thesis, development of a safety system for the prevention of collisions between electric wheelchair and surrounding obstacles is presented. The existing solutions and theoretical background of the operation and the scope of use of light and ultrasonic sensors for distance measuring are presented. After performing a comparison between the sensors in terms of accuracy, ultrasonic sensors are selected, because their value of the measured distance is the most similar to the actual one. To provide physical connections between the external control module of the wheelchair, arrow keys, the Arduino Micro microcontroller and the existing controller, which allows operating with the tongue, a printed circuit board is developed with the associated connecting cables. A development of the control algorithm as a result of decision rules is determined by the machine learning technique. On the basis of the measured distances from all sensors, the microcontroller determines and simultaneously graphically displays on the screen in which direction the movement is permitted and in which is not.