This paper presents a novel method for estimating the parameters of electro-mechanical systems, or known as actuators, within limited stroke. Identification of actuators with unlimited stroke is an ordinary problem in control applications. However, in identifying parameters of the actuator like the viscous damping coefficient, the Coulomb friction torque, and the equivalent moment of inertia reflected to the motor side, the mechanism with limited stroke is involved and the identification problem becomes more complicated. In this paper, a simple and straightforward method is developed to resolve this problem. A sequence of pulses with various levels is designed to estimate the viscous damping coefficient and the Coulomb friction torque within the limited stroke of the actuator. Then an optimal algorithm based on the interior-reflective Newton method is applied to search the moment of inertia of the overall system. It is demonstrated in the simulation examples that even when the velocity profile is calculated directly from the first difference of the sampled position and is corrupted by highly amplified noise, the proposed method can still find the moment of inertia with error less than 1%. Simulation and experimental results are both presented indicating the practical viability of the proposed approach.