Hand gesture recognition has become a major focus of research in the field of human-computer interaction (HCI). This work proposes a static hand gesture recognition system. The Histogram of Oriented Gradients (HOG) was used for feature extraction. The features are reduced by PCA and further reduced using the attribute reduction algorithm in the theory of the neighborhood rough set. Then, the weight of every feature is calculated using the attribute significance algorithm in the theory of neighborhood rough sets. The weighted features are applied as input to the fuzzy neural network to recognize static hand gestures. Experimental results on commonly-referred databases show that the proposed method based on neighborhood rough sets improves the recognition accuracy of fuzzy neural networks.