Haptic robots are used for simulating virtual objects. Since most real objects can be modeled with a spring and damper, virtual objects are also implemented as a spring and damper (in a discrete space). Stable performance of the haptic robot results in convergent vibrations, while unstable performance leads to divergent vibrations. On one hand, the haptic robot should have negligible effective mass and friction so that its dynamics can be disregarded in comparison to the virtual object’s dynamics. On the other hand, the dynamics of the user’s hand affect the stability and performance of the robot. In this study, the dynamic parameters of the user’s hand on the haptic robot were identified and optimized using the Artificial Bee Colony method. First, the theoretical stability boundary was presented as a function of the dynamic parameters of the user’s hand on the haptic robot, the stiffness and damping coefficient of the virtual object, sampling time, and time delay. Concurrently with experiments involving the user’s hand placed on the lightweight KUKA 4 robot at the PRISMA lab in Italy, the empirical stability boundary was obtained. Then, using the Artificial Bee Colony optimization method, the dynamic parameters of the user’s hand on the haptic robot were determined such that the error between the theoretical and empirical stability boundaries was minimized. The evaluations indicated a high accuracy of this method in determining these parameters and consequently the stability boundary of the robot.