Abstract: To overcome the limitations of existing gas pipeline leakage location technologies-such as low accuracy, poor adaptability to operating conditions, and insufficient capability to detect micro-leakages-this study proposes a precise leakage location method based on a multi-source signal attenuation model. In this method, monitoring nodes are arranged using the triangular coverage principle to synchronously collect pressure wave, acoustic wave, and optical fiber vibration signals, followed by targeted denoising. Separate signal attenuation sub-models, incorporating factors like gas viscosity, density, pipe diameter, and main acoustic frequency, are fused into a unified attenuation model using dynamic weights determined by information entropy. The gradient descent algorithm calculates the leakage point coordinates, while attenuation coefficients are optimized through a closed-loop verification mechanism. Experimental results indicate that the method limits the location error to under 5 m and detects micro-leakages below 0.1 m³/h. It applies to medium-, low-, and high-pressure pipelines transporting gases such as natural gas and LPG, maintaining high stability even under intense interference with a 10 dB signal-to-noise ratio(SNR). Operation and maintenance costs are reduced by over 20% compared to existing technologies. This method offers reliable technical support for daily operation and emergency response in gas pipeline networks.