Abstract: The key problem that the UWB in the underground imaging algorithm needs to be solved is to image the tar- get buried on the other side of the cave body with the unknown electromagnetic parameters. It is difficult to effectively predict and analyze the performance of UWB detection imaging system in mine because of the influence of physical phenomena such as complex electromagnetic properties and multipath propagation effects in the backlash of roadway. Therefore,a method to combine a ray tracing algorithm (Ray-tracing) and SVM was proposed. This method uses Ray- tracing algorithm to obtain the sample data of penetrating the image of the roof collapse,and then classifies it with SVM to solve the problem of buried target recognition after the roof collapse of the mine. The simulation results show that the method can eliminate the non-linearity and ill-posedness in the inversion process,and can realize the unknown target imaging after the roof collapse.