This paper focuses on the forecast of wind shear and turbulence at the Hong Kong International Airport. It presents a mesoscale prediction model that uses chaotic oscillatory-based neural networks (CONN) to forecast the evolution of wind fields along the glide path in the vicinity of the airport. This model makes use of accurate Doppler velocities measured by light detection and ranging (LiDAR) system and afterward collected by the Hong Kong Observatory. Simulation results show that the CONN model with a new learning algorithm is able to capture the occurrence, evolution, and sudden changes of the winds representing turbulence incidences in the region. Research findings show that Doppler velocities forecast using CONN can be transformed into headwind profiles and processed with the developed algorithm to identify the wind shear occurrence. These are shown to match actual observations made using LiDAR in terms of time, locations, and size of wind shear events with considerable accuracy. The model has better performance compared with that of the traditional multilayered perceptron model neural network. The results encourage further exploration and experimentation in the use of machine learning and chaotic neural network in weather forecast and alerting.