 This study investigates the application of two artificial neural network, ANN, based on symbol methods, multi-boost ANN, and rotation ANN, for wetland cover classification in the Zoage wetland on the King Heiterbet plateau. The results show that both methods significantly improve single ANN, and outperform other classifiers, such as VGG-11, and random forests additionally, the IANN, and MAN are found to be more robust to data size reduction, feature variability, and noise. Overall, Ensemble Learning provides a promising scheme for wetland cover classification. This article was authored by Su Dong-hu, Peng Lin-jong, Qi Zhong, and others.