This paper presents the application of an output feedback Nonlinear Receding Horizon control algorithm to a laboratory seesaw equipment. This control law guarantees exponential stability of the equilibrium and allows one to consider the presence of control and state constraints. Since the specific control application requires a small sampling interval (0.05s), the nonlinear control law is computed off-line for different values of the initial state. Then, an approximating function is derived with the aid of a Neural Net, which is subsequently implemented for on-line computations.