Turbochargers are most widely used in automotive, marine and locomotive applications with diesel engines. To increase the engine performance nowadays, in aerospace applications also turbochargers are used. Mostly the turbocharger rotors are supported over the fluid film bearings. With the operation, lubricant properties continuously alter leading to different load bearing capacities. This paper deals with the diagnostic approach for prediction of shaft unbalance and the bearing parameters using the measured frequency responses at the bearing locations. After validating the natural frequencies of the rotor finite element model with experimental analysis, the response histories of the rotor are recorded. The influence of the parameters such as bearing clearance, oil viscosity and casing stiffness on the unbalance response is studied. By considering three levels each for shaft unbalance and oil viscosity, the output data in terms of four statistical parameters of equivalent Hilbert envelopes in the frequency domain are measured. The data are inversely trained using Radial Basis Function (RBF) neural network model to predict the unbalance and oil viscosity indices from given output response characteristics. The outputs of the RBF model are validated thoroughly. This approach finds changes in the rotor bearing parameters from the measured responses in a dynamic manner. The results indicate that there is an appreciable effect of lubricant viscosity at two different temperatures compared to other parameters within the operating speed range. The identification methodology using the neural network is quite fast and reliable. © 2019, Shahid Rajaee Teacher Tarining University (SRTTU). All rights reserved.