The system uses digital camera to collect video images of cockpit real-time,applies the technology of digital image recognition to find which component that the user is interested in,then displays the corresponding training information on HMD(head-mounted-display).The design of this system includes the parts of hardware and software.The design of hardware is integrating USB digital camera,mini HMD,wearable computer and assistant power module into a moveable augmented reality system;the design of software consists of the creation of the cockpit template library,the creation of the aided information library and the implementation of SURF(speeded up robust features) image feature match algorithm based on OpenCV(intel open source computer vision library).The experiment demonstrates this system can provide the training information real-time for the user to accomplish the training tasks.