Abstarct: There are several ways for modeling a system. Modeling is usually applied using physical relationships governing the system or system identification methods or a combination of both. In general, these models may be static or dynamic, linear or nonlinear, time variant or time invariant, continuous-time or discrete-time, deterministic or stochastic. In one of these methods which is called black box modeling, at the beginning a test is performed for modeling the system and input and output data are recorded and then this data is used for identification and modeling of the intended system. In this thesis a Phantom robot has been identified using neural network. To evaluate the performance of the proposed methods, a Phantom robot has been chosen and experimentally identified using data extracted from the robot. The Phantom robot has three manipulators and is widely used in haptic and telerobotic applications. Modeling the robot, is important for predictive applications, simulation, control, and fault detection. The results show that neural networks are able to identify a Phantom robot with high precision.