The Artificial Neural Network (ANN) and curve fitting are traditionally prevalent for prognosis of bearing due to the simplicity of application. The fusion techniques are known to increase the diagnostic capabilities of the damage identification parameters (such as RMS, Peak etc.). However, the role of the fusion techniques in the prognosis of a bearing is still to be explored (attempted in this work). The Mahalanobis-Taguchi-Gram-Schmidt (MTGS) technique is used to fuse various damage identification parameters into a single fused parameter, Mahalanobis Distance (MD). The MD is used as a damage identification parameter for the ANN and the curve fitting. Three methods (i) ANN with damage identification parameters (ii) ANN with MD and (iii) Curve fitting with MD are compared for the effectiveness of the prognosis. The polynomial fit, the Gaussian fit, the sum of sinusoidal fits, and the sigmoid shape fit are attempted. The vibration data for these methods is acquired from a naturally induced and progressed defect through an accelerated life test. © Proceedings of ISMA 2018 - International Conference on Noise and Vibration Engineering and USD 2018 - International Conference on Uncertainty in Structural Dynamics. All rights reserved.