 This study presents a random forest classification approach to retrieve wetland land cover in arid regions by fusing Pleiad-1b data with multidate Landsat-8 data. The segmentation of the Pleiad-1b multispectral image data was performed using an object-oriented approach, and geometric and spectral features were extracted for the segmented image objects. NDVI series data were also calculated from the multidate Landsat-8 data to reflect vegetation phenological changes in its growth cycle. The feature set extracted from both sensors data was optimized and employed to create a random forest model for wetland land cover classification in the Ertix River in northern Xinjiang, China. Comparison with other classifiers such as support vector machine and artificial neural network indicates that the random forest classifier can achieve accurate classification with an overall accuracy of 93% and kappa coefficient of 0.92. The classification accuracy of farming lands and water bodies was improved by incorporating geometric shapes, phenological difference, and textural information of co-occurrence gray matrix. The inclusion of phenological information in the classification reduced errors and improved overall accuracy approximately 10%. The results show that the proposed random forest classification by fusing multisensor data can retrieve better wetland land cover information than other classifiers, which is significant for monitoring and managing wetland ecological resources in arid areas. This article was authored by Xiaohang Tian, Xinfeng Zhang, Jia Tian, and others. We are article.tv, links in the description below.