To improve the predictive performance, applicability, and accuracy of the civil aviation wind shear risk prediction model, a whale optimization algorithm (WOA) with dynamic adjustment strategy and constraints was proposed to optimize the gated recurrent unit (GRU) of the civil aviation wind shear risk prediction model. Firstly, based on the data of civil aviation unsafe incidents, a wind shear risk index system was constructed and a quantitative calculation method for monthly average risk index was proposed, providing standardized input for wind shear risk prediction. Secondly, to address the problem of traditional WOA optimization of high-dimensional parameters being prone to local optima or low search efficiency, a hyperparameter sensitivity factor was introduced to make the model approach the global optimal solution faster. Thirdly, utilizing the advantages of WOA global search and GRU temporal feature extraction, a WOA-GRU combination model was constructed and applied to the field of civil aviation wind shear risk prediction. The results show that compared with back propagation(BP) neural network and convolutional neural network (CNN) models, WOA-GRU reduces mean square error (MSE) by 65.45% and 74.91%, respectively, and reduces mean absolute error (MAE) by 64.66% and 65.85%, respectively. The prediction accuracy performance of the model is better than other reference models. The model fits the historical sequence of monthly wind shear risk well, demonstrating higher accuracy and reliability in wind shear risk prediction and warning.