Condition monitoring techniques provide vital data for operators to avoid unpredicted and unwanted stops of machines caused by faults. One of these techniques is vibration analysis, which is used for faults diagnosis and prognosis such as shaft bending, misalignment, lousy bearing, worn gears, unbalances of rotors, etc. Moreover, vibration signals can be employed in intelligent algorithms like Fuzzy Models, Support Vector Machines, and Neural Networks to prepare better and more accurate predictions of current and future conditions of the machine. This paper discusses the application of vibration signals in the prediction of rotor unbalance parameters including the unbalance location and amount. Some statistical features were applied to the inputs of the neural network that had been derived from the time and frequency domains of bearing acceleration signals. The experimental study shows that the developed model can estimate these parameters with acceptable accuracy.