In last decades, Gabor feature based face representation presented promising results in face recognition applications due to its robustness against illumination and facial expression changes. The power of Gabor lays in its properties like the computation of local structure corresponding to different spatial frequency (scale), spatial localization, orientations and inessentiality of manual annotations. This work, an Ensemble based Gabor Nearest Neighbor Classifier (EGNNC), extends the Gabor Nearest Neighbor Classifier (GNNC), which extracts important discriminative features utilizing both the Gabor filter and Nearest Neighbor Discriminant Analysis (NNDA). EGNNC is an ensemble classifier combining multiple NNDA based component classifiers which are designed using different segments of the reduced Gabor features. EGNNC has better use of the discriminability implied in reduced Gabor features by avoiding 3S (small sample size) problem in contrast to GNNC, since GNNC operates on the whole Gabor feature space and reduces the dimension by one component NNDA. In experimental analysis EGGNC outperformed its ancestor GNNC in terms of recognition rate and accuracy for the two test datasets which makes the proposed ensemble approach applicable in face recognition.