We employ the pretrained deep CNN models for VHR images scene classification, where we used VGG-Net as the feature extractor by selecting useful layers in order to get a good representation of the images scene.