The efficient separation of coal and gangue is an important way to realize green mining of coal resources, and the core technology is the rapid and accurate identification of coal and gangue. Therefore, a coal-gangue recognition method based on the fusion of X-ray image and laser point cloud is proposed in this article. Firstly, an improved Otsu segmentation algorithm based on the local entropy and global mean difference weighting is designed to enhance the segmentation accuracy and efficiency of X-ray images. Meanwhile, the straight-through filtering and voxel grid down sampling are used to simplify the laser point cloud data of coal and gangue, and the coalgangue feature combination of X-ray image and laser point cloud is extracted. Then, to address the problems that the traditional sparrow search algorithm (SSA) is prone to fall into local optimum and the population diversity is poor, a multi-strategy improved SSA algorithm (ISSA) is proposed to optimize the model parameters of light gradient boosting machine (LightGBM). A coal-gangue fast recognition model based on ISSA-LightGBM is designed. Finally, an experimental platform for the coal-gangue recognition is established and the corresponding experimental comparative analysis is carried out. Results show that the comprehensive recognition accuracy of ISSALightGBM model can reach to 99. 00% , and the comprehensive performance is superior to other models, which could meet the needs of efficient coal-gangue recognition.