This paper presents a robotics vision-based system for an underwater pipeline and cable tracker. This system is introduced to improve the level of automation of underwater remote operated vehicles (ROVs) operations which combines computer vision with an underwater robotics system to perform target tracking and intelligent navigation. This study focuses on developing image processing algorithms and nonlinear control method for tracking the pipeline or cable. By using the proposed image processing method, at first the picture frames are converted to gray pictures and these images should be applied for an edge detector at last by using Hough transform the angle of the robot can be found. This angle is used as a reference input of heading controller. By using the adaptive sliding mode controller, the ROV can easily track the cable or pipeline. The system has been tested using sequences coming from a video tape obtained in several tracking sessions of various cables and pipelines with an ROV.