In the disease developing mechanism of baseball elbow, it is believed that there is a need to understand the skeletal system of the elbow joint and forearm. Focusing on the interior of a elbow joint, the humerus, ulna and radius are constituted a complex structure covered with soft tissue, such as the joint capsule and collateral ligaments. In order to clarify the failure of the forearm, Kecskemethy et al., considered a simple forearm skeleton model. Although they estimated ulnar behavior by adjusting the stiffness of the model, Nojiri et al. proposed a method for estimating the link length and the measurement error gain using the steepest descent method (SDM) as another approach. However, since the least squares method (LSM) has a possibility of falling into a local solution; hence, in this paper, we propose a method for estimating in Particle Swarm Optimization (PSO). In addition, by estimating the link length and the measurement error gain that made dependent on the pronation supination (pro-/supination) posture, we indicate that was able to carry out the reduction of estimation error.