For the application Of augmented reality in plant maintenance work, real-time tracking and technology with higher accuracy is necessary. This Study focuses oil the tracking method in vision based SLAM. In NPP, line features are abundant, and they are detected more easily and reliably than point features. Line features offer more information than points, but its tracking method is more complex. In this Study, line features are used as landmark for tracking The representation of the 3D line is relied on Plucker coordinates. A Gaussian sum approximates the feature initial state and is updated as new observations are gathered by the camera. Then extend Kalman filter is adopted for SLAM approach.