In spite of advanced electro-mechanical technology on passenger or load elevators, elevator accidents still occur. Therefore, it is necessary to analyze vibrations of elevators with and without load for predicting some possible faults on their mechanical parts. This study proposes an adaptive neural network predictor to estimate and evaluate the vibrations on elevator systems. For this purpose, elevator vibrations are measured from two points of the elevator system for different working conditions, and different types of neural network analyzers are employed to evaluate the system vibrations. Simulation results show that neural networks can be used as an adaptive analyzer for such systems in the experimental applications.