The thesis describes the process of optimizing the trajectory of a mobile robot using artificial intelligence. The basic concepts of robotics are explained, such as joints, end-effectors and sensors with which the robot performs tasks. The construction and ways of manipulating the mobile robot in the environment within the CoppeliaSim software package have been elaborated. Then, some artificial intelligence methods are described that can optimize the path of a mobile robot through a maze, from which Q-learning is chosen. Maze generation models using the Depth-first Search method, as well as a Q-agent training model, were developed in the Python programming language. Using the graphical user interface, the results of the optimization of the mobile robot's path were obtained in graphic form. Using these results, a scene was recreated in the CoppeliaSim environment by creating a maze model and running a mobile robot through that maze. The mobile robot reached the goal through the optimal path retrieved from the Q-learning model, and using the logic explained in the paper. The result was compared with a primitive algorithm, and it is obvious that the length of the path with the application of artificial intelligence is significantly shorter, that is, optimal.