In this paper, a RQA-based approach is developed for feature generation from raw vibration data recorded from a rotating machine with five different conditions. The created features are then used as the inputs to a classifier for the identification of six bearing conditions. Experimental results demonstrate the ability of RQA to discover automatically the different bearing conditions using features expressed in the form of recurrence quantification measures. Furthermore, using RQA extracted features and traditional features with artificial neural networks (ANN) and support vector machines (SVM) have been obtained. This RQA-based approach is used for bearing fault classification for the first time and exhibits superior performance over other traditional methods.