Aircraft often operate in complex and variable dynamic load environment， and dynamic load localization is the primary problem that needs to be solved in this field. This paper focuses on the dynamic load localization requirements of common and prone to abnormal vibrations in aircraft structures. Combining deep neural network， a rapid dynamic load localization method for aircraft structures is developed. By using Long Short-Term Memory （LSTM） neural network， the inverse implicit function model which can accurately describe the corresponding relationship between the dynamic load location and vibration response of the struc? ture is constructed. A dynamic load localization method based on the LSTM neural network classification model is proposed. A sim? plified finite element model of the entire aircraft structure is established to simulate several typical dynamic load conditions that the aircraft may encounter during actual flight. The noise resistance and robustness of the established deep neural network are also studied. The simulation results show that the proposed method can accurately identify the location of dynamic loads under various load condi? tions， and can still maintain high locating accuracy under the measurement noise level of 10 dB and the parameter perturbation of 2.8%.