Subsea pipelines and cables are critical assets which require regular maintenance and inspection to ensure their integrity and continual operation. The autonomous tracking of these assets requires robust and reliable methods especially in the challenging subsea environment. This paper presents a new method for the robust autonomous detection and tracking of subsea pipelines and cables using a multi-beam echo-sounder sensor, leveraging intensity and profiling returns for enhanced robustness. The proposed method involves four key steps. First, prepocessing operations are carried out to refine the raw sensor data, followed by a region of interest generation using the K-means clustering algorithm, then a validation step which filters implausable regions and finally a fitting processes for determining the target’s position and parameters. The proposed method is also designed to extend the detection and tracking capabilities of the system to the 3-dimensional use case. Through real-world and simulated experiments we demonstrate the effectiveness of the method.