Servo-actuated mechanisms are increasingly substituting fully mechanical drives in order to increase flexibility and reconfigurability of modern automatic machines. The overall servomechanism performance, especially in the case of high-dynamic motions, is the direct consequence of several interacting factors, namely electric motor and linkage dynamics, controller efficacy, and requested motion law. In particular, Point-To-Point (PTP) trajectories are usually designed in order to comply with technological constraints, imposed by the required interaction with the handled product, and to maximize some optimality criterion such as, for instance, energy efficiency or limited actuation torques. In this context, the present paper proposes a novel method for generating either energy-optimal or torque-optimal PTP motions described by piecewise fifth-order polynomials. The optimization cost functions are based on a virtual prototype of the system, which comprises behavioral models of power converter, controller and electric motor coupled with the mechanical system. Results are then compared with experimental data obtained on a physical prototype. The comparison quantitatively shows that betterbehaved PTP trajectories can be designed by including the dynamic contribution of each sub-system component.