In this study, cascade PD+P controller gains for the ball and beam system are tuned using metaheuristic algorithms. The aim of the ball and beam system, which provides the simulation of many systems in the laboratory environment, is to reach the reference position and balance of the ball, which moves freely on a single axis on the beam. PD+P cascade controller which contained fast and slow dynamics at the same time is designed for the ball and beam system. Cascade controller gains, which are difficult to adjust compared to conventional controllers, are tuned using Artificial Bee Colony and Teaching Learning Based Optimization algorithms. In the optimizations applied, IAE, ITAE, ISE, MSE objective functions, and a new step response based objective function are proposed. In the controller responses obtained using the proposed objective function, the ball reached the desired reference position without overshoot.