We evaluated object detection and classification with geometric context information on three object categories from dataset of PASCAL VOC 2010. They are plane, car and person. The accuracy of detection is measured in term of Average Precision (AP). The object categories detection without context is implemented with the discriminatively trained part based models [1]. We take training dataset of PASCAL VOC 2010 as training set. The experiments of object categories detection are conducted on validation dataset of PASCAL VOC 2010.