Abstract: Conventional seismic attribute interpretation method uses single seismic attribute, or multiple seismic attributes to separately interpret and verify each other to identify subsurface anomalous geological bodies. Although it can meet the exploration requirements to a certain extent, the accuracy is unsatisfactory. In order to accurately identify the distribution and boundary of the coal seamroadway, an interpretation method based on seismic multi-attribute fusion technology is proposed. A geological model was designed, and itsseismic response characteristics were studied. Aroadway can be regarded as a special geological hollow body that is in the coal seam and has the common top and bottom rock formation with the coal seam. Generally, the roadway is filled with water or air. The strong reflection interface formed due to the obvious impedance difference between the roadway and the surrounding rock will lead to the phenomenon that the reflected wave energy at the roadway becomes stronger locally. Extracting a variety of seismic attributes from the forward results showed that the relative acoustic impedance and amplitude attributes can well describe the seismic reflection anomalies caused by the roadway. Three different seismic attributes including relative acoustic impedance, instantaneous amplitude and mean amplitude were extracted from the seismic volume along the target coal seam. These three seismic attributes were regarded as R, G, and B color components, respectively, and fused by RGB color fusion technology to obtain afusion attribute including three seismic attribute information. The position and boundary of the coal seam roadway wasfound out successfully by using the fusedattributes, which showed that it is effective to use seismic multi-attribute fusion technology to identify roadway. The results showed that, compared with the single seismic attribute, the attribute after the fusion of multiple seismic attributes can show more geological information, and the roadway and its boundary were clearer and more continuous, which improved the interpretation accuracy and provided a basis for detecting and identifying coal mine roadway.