Abstract: To address the limited adaptability of traditional static scoring systems (such as the Kent method) and the lack of quantified multi-factor coupling in the GB/T 27512—2011 Risk Assessment Methods for Buried Steel Pipelines standard, a natural gas pipeline risk assessment model with dynamic correction and risk coupling was developed to enable more precise classification of failure probabilities and consequences. A five-dimensional failure probability indicator system (corrosion, design, third-party damage, misoperation, natural disaster) was established, with the basic failure probability calculated from historical accident data and dynamically adjusted using the correction coefficient η_i in conjunction with actual pipeline parameters. A medium-receptor double-layer coupling consequence assessment system was constructed, in which the coupling matrix was applied to quantify the interaction between medium hazard and receptor sensitivity. The probability and consequence levels were integrated through a risk matrix, and comprehensive risk quantification results for levels Ⅰ-Ⅴ were produced. The model was applied to a natural gas pipeline in China, revealing a comprehensive failure probability of 8.781×10^-5/(km·a), with design defects contributing the most (34.2%). The failure consequence score was 792 points, indicating minor risk, while medium diffusivity significantly affected population density. The overall risk level was classified as Grade Ⅰ (minor), with operating pressure and soil corrosivity identified as key risk factors. The model not only directly outputs a quantitative failure probability but also accurately identifies key risk transmission paths. It supports precise prevention and control in high-consequence areas and advances pipeline risk management from static to intelligent dynamic approaches.