Condition monitoring is an important and effective machine maintenance strategy which emphasize to carry out the maintenance work only, when the condition of machine or equipment demands so for safeguarding it from incipient failure. It is an effective maintenance tool for reducing the maintenance cost, machine down time and preventing the unscheduled breakdown or shutdown of the machine, thus increasing the plant availability, and ensuring overall safety of men and machines Recent years in the condition monitoring of machineries various neural network models have been applied successfully for the detection and diagnosis of machinery faults. Present paper highlights the condition monitoring studies on an Industrial Fan discussing the vibration monitoring and data preparation procedure, which have been used as input to artificial neural network models, thereafter the feasibility study based upon training and testing of Back-propagation (BPNN) and Radial basis function neural network for detecting and quantifying the fault.