As a key technology to ensure the safety of energy transmission and fluid transmission, its development is of great significance to improving the reliability of pipeline systems and promoting industrial intelligence. This paper systematically reviews the research progress and challenges of pipeline robots in the fields of mechanism design, driving methods, motion control, and detection technology. The mechanism design shows a diversified trend: wheeled drive is efficient but not adaptable to complex environments. The Tracked drive has strong barrier-over ability and high energy consumption. Spiral drive is suitable for large pipe diameters but is susceptible to media viscosity. Peristaltic damage is small but has limited speed. Foot type is good and has complex control. Soft-bodied actuation has outstanding flexibility but weak load-bearing capacity. Motion control technology is developing towards intelligence. Traditional proportional-integral-derivative (PID) control is combined with intelligent algorithms (such as neural networks and reinforcement learning) to improve positioning accuracy through multi-sensor fusion, but real-time and complex fluid-solid coupling modeling are still difficult. The detection technology combines multi-source sensing, such as magnetic leakage, ultrasound, eddy current, and vision, and combines deep learning to achieve intelligent defect recognition, but it cannot detect and predict the lifespan of micro defects. Application scenarios extend from oil and gas, drainage pipelines, to special environments such as seabed and electricity, and face adaptability challenges such as high-pressure sealing and grain subsidence. Future research needs to focus on rigid-flexible coupled driving structure, data-model hybrid control framework, multi-physics collaborative detection system, and deepen the adaptability of special working conditions and multi-equipment collaboration mechanisms to break through the limits of pipe diameter adaptation, improve independent operation capabilities, and engineering application level.