Trajectory (path) planning is a well known and thoroughly studied field of automated planning. It is usually used in computer games, robotics or autonomous agent simulations. Grids are often used for regular discretization of continuous space. Many methods exist for trajectory (path) planning on grids, we address the well known A* algorithm and the state-of-the-art Theta* algorithm. Theta* algorithm, as opposed to A*, provides 'any-angle' paths that look more realistic. In this paper, we provide an extension of both these algorithms to enable support for speed limit constraints. We experimentally evaluate and thoroughly discuss how the extensions affect the planning process showing reasonability and justification of our approach.