Overview: Wake vortices develop as a consequence of the lift an aircraft produced to fly. For a wing generating lift, the pressure on the wing lower surface is higher than the pressure on the wing upper face. Therefore, air flows around the wing tip from the lower surface to the upper surface resulting in a strong vortex, the so-called "wing tip vortex". An airplane affected by a wake vortex experience may cause rolling moment even air crash. Given that, how to recognize wake vortex and monitor it to improve the capacity of airdrome and airspace, has become a key issue in civil aviation industry. The traditional method of detecting wake vortex generally adopts Doppler LiDAR, which is considered one of the most effective approach. In this paper, the LiDAR made use of the range-height-indication mode to obtain the radial velocity of the wake vortex, and the tangential velocity was calculated by the Hallock-Burnham vortex model, and then converted the velocity data to the speed maps of the vortex through processing. With the rapid development of artificial intelligence, convolution neural networks has turned out to be a powerful tool to deal with image analysis. For this reason, this paper applied AlexNet neural network model combined with the detection principle of Doppler LiDAR to extract the image features of the wake vortex velocity images in the atmosphere and identified the aircraft wake vortex by training a large amount of vortex maps. Aiming at perfecting the data sets, this experiment collected the flight departure data within 40 days of an airport in China. The airport took off about 500 flights a day, including A340, A380 and ARJ21 and so on. The AlexNet was trained and tested on the designed data sets, which involved 4000 training sets and 1000 validation sets and the training epochs were set as 10000. The qualitative experiment results show that after the network model converges, the accuracy rate reaches to 91.30%, which can effectively realize the identification work, monitoring of the aircraft wake vortex, as well as early warning. This research demonstrates the high accuracy and low probability of false alarm of the AlexNet neural network in detecting wake vortex and is capable to provide decision-making information for air traffic control work.