Paper presents effect of load conditions onresults of feature extraction technique used for outerrace bearing fault detection system in an inductionmotor. Induction motor, as a widely used machine inthe electrical field, its overall life is based on itstimely identification of fault scenario. Fastestappropriate corrective action can be taken on thebasis of fault analysis. Various extraction techniquesare implemented for finding faults observed in theinduction motor. Now days it is common practice totake corrective action using digital signal processingtechnique. In this paper the fault signature is carriedout by comparing two bearings, mainly healthy andfaulty bearing, mounted on the same machine. Database is obtained for stator current for healthy andfaulty bearing operated at different load conditions.The readings obtained for healthy and faulty bearingdiffer from each other. But the actual currentsignature observed is very small and need to beextracted very precisely. The technique of faultextraction using wavelet transform is already used insome cases earlier. This paper focuses on observedchange in the fault signature with respect to variousload conditions of the motor. The result analysisshows that the fault signature is obtained more orless similar in all the load conditions and can bedetected bearing fault precisely.