In the mining and underground engineering, it is necessary to accurately obtain the hydrogeological parameters of the aquifer in order to do a good job in water inflow prediction and water disaster prevention. Based on this, the particle swarm optimization algorithm ( PSO ) is used to study the engineering practice of hydrogeological parameter identification and groundwater level prediction, and the coupling technology of PSO and Theis formula and the factors affecting the accuracy of parameter inversion are mainly introduced. The effectiveness of this method is proved by typical pumping test. Combined with engineering practice, the application of this technology is expounded, and the good benefits obtained from the optimization of drainage scheme, evaluation of groundwater resources and dynamic prediction of water level in mining area are discussed. The problems existing in this technology under complex recharge conditions and strong heterogeneous aquifer conditions are discussed. Finally, the future work direction is put forward, that is, intelligent optimization of algorithm parameters, establishment of test database under different backgrounds, and improvement of various physical field coupling prediction models, so as to improve the adaptability and prediction reliability of this technology to complex hydrogeological conditions.