Abstract: Among the modern technologies used in air traffic surveillance systems, Automatic Dependent Surveillance-Broadcast (ADS-B) is the most eye-catching one today, with higher accuracy and less human dependence. However, it broadcast messages without any authentication and encryption, and the information may be maliciously forged or modified. This article introduced two network models based on 1DCNN-BiLSTM. This model extracted the characteristics of the real signal and the spoofed signal based on the ADS-B time-domain sampling data and identified the spoofed signal. When the trajectory was short, the model could extract the detailed time structure information of the ADS-B signal; when the trajectory was long, first used one-dimensional Convolutional Neural Network (1D-CNN) to extract each air position in each trajectory, Then used the Bi-directional Long Short-Term Memory (BiLSTM) to mine the spatial relationship between different air positions. After simulation experiments, the 1DCNN-BiLSTM network has a better detection effect than the network that only has time feature extraction, such as DNN and LSTM.