In recent years, some object detection algorithms have been proposed for X-ray image testing. Most of them are developed for baggage inspection to detect target objects automatically. However, these automated detection algorithms are still very limited to meet the need of practical application and X-ray screening systems are still being manipulated by human inspectors mainly. This situation of inspection is very unsuitable for parcel inspection in logistics industry owing to the huge amount of parcels. In this paper, we propose a classification method for X-ray image, which is a new method for automated inspection and can promote the inspection efficiency of logistics industry effectively.