Abstract: In the tunnel and coal mine underground roadway construction process, serious hydrogeological disasters are faced. Under the complicated construction environment and hydrogeological conditions, it is very difficult to accurately predict water damage. As a widely used electromagnetic induction exploration method, the mine transient electromagnetic method is mainly used to detect the water damage as the water-bearing stratum which is low in electrical conductivity. And the mine transient electromagnetic method is sensitive to low resistive bodies and is capable of detecting the aquifers in the tunnel and mine roof and floor. In order to solve the problem that the data processing of roadway or tunnel transient electromagnetic detection is difficult and the accuracy is not high, the two algorithms of particle swarm optimization (PSO) and damped least squares (DLS) are compared and analyzed, which solves the problems that PSO algorithm has lower optimization efficiency in the later stage and DLS algorithm requires manual assignment of initial values. With reference to the actual stratum data, the mathematical model was established, and the inversion trial calculation was carried out by using PSO, DLS and the combined algorithm was proposed in this paper. The results showed that the combined algorithm had higher optimization efficiency and did not need to manually give the initial model, comparing with the two separate algorithms. The combined algorithm was used to calculate the measured transient electromagnetic data of the mine roadway in Huaibei Coalfield, China. The calculation resulted in anomalous separation of the top and bottom plates. The interpretation results were consistent with the logging data, geological data and roadway exposure conditions. It proves that the combined algorithm can invert the roadway transient electromagnetic detection data with high accuracy.