Abstract: The evaporation duct is highly available in the maritime combat environment, and it is an important factor affecting the performance of shore-based, shipborne, or low-altitude airborne radars and communication system. This paper uses the Elman neural network is used to construct an intelligent prediction model of evaporation duct characteristics based on meteorological data and to realize the precise prediction of the characteristic parameters of the height of evaporation duct. The sea experimental sounding data are collected in the paper and the constructed model is used for the prediction of the height of evaporation duct. The analysis results show that the model can deeply learn or dig the spatiotemporal elements from the meteorological parameters according to sea observation data. The prediction error of all 476 groups of data is 1.94 m, and the statistical results are in line with the normal distribution law. The prediction error of the test set is 1.90 m, indicating that the intelligent prediction model of evaporation duct characteristics has a good generalization capability in prediction. It can provide favorable support to occupy the advantage of the electromagnetic environment and spectrum warfare in maritime combat environment.