Abstract: Wheat is one of the most important crops for many countries. It is important to obtain accurate wheat planting area. Optical imagery has been widely used in crop classification and made great progress, but the optical imagery in the critical period of wheat growth is often affected by weather and other external conditions. In contrast, synthetic aperture radar (SAR) is potentially effective data source for wheat mapping because of its all-weather, all-day imaging capabilities. The study site was located in a typical urban agricultural region in Gaochun District of Nanjing, the provincial capital of Jiangsu Province, China, with the central coordinates of 118°52'E and 31°19'N. Six SAR (Sentinel-1A) and 4 optical (Landsat-8) images were obtained from the ESA during the growing period of winter wheat. The objective of this study was to explore the abilities of multi-temporal satellite SAR data, optical images and the combination of both types of images to identify winter wheat in an urban agriculture region with complex planting structures. The investigations were through the following steps: 1) The texture features were extracted from the SAR images using the gray level co-occurrence matrix (GLCM), and the backscatter intensity data were also obtained; 2) A support vector machine (SVM) was used as the classifier to map winter wheat, using the different combinations of the Sentinel-1A derived information and optical images. 3) Then, the classification accuracies for different combinations of SAR variables and optical images were evaluated. The results showed that: 1) In the process of winter wheat growth, the backscatter characteristics changed obviously. In the middle of April, the backscatter characteristics were different between winter wheat and other vegetation in the study area, which could provide the theoretical basis for winter wheat classification. 2) Using single polarization, single-temporal SAR data were difficult to meet the accuracy requirements of winter wheat with complex planting structures, and the classification accuracy of VV polarization was higher than that of VH polarization; 3) Compared with VV polarization, the producer's and user's accuracy of VV+VH combination were increased by 27.3% and 13.6%, respectively. When VV polarization was added to VH backscatter intensity images, the producer's and user's accuracy were increased from 50.3% to 85.7% and from 54.5% to 87.9%, respectively; 4) When texture images were added to backscatter intensity images, the winter wheat fields could be mapped with a satisfactory classification accuracy, and the user's and producer's accuracy could reach up to 94.6% and 93.3%, respectively; 5) The combination of multi-temporal SAR and optical images had the best classification accuracy, and the producer's and user's accuracy were 97% and 98.4% (overall accuracy 95.3%, Kappa coefficient 0.91), respectively. The addition of multi-temporal SAR to optical images allowed us to increase producer's and user's accuracy of winter wheat by 3.7% and 3.8%, respectively. The results in this paper show that the satisfactory classification of winter wheat can be achieved using Sentinel-1A data alone in urban agricultural regions with complex planting structures. In addition, texture features of SAR can improve SAR data classification accuracy. The SAR and optical images can make use of the advantages of information complementation of each other to make up for their deficiencies in crop classification. It is indicated that the SAR data can not only be used as a data source instead of optical images to distinguish winter wheat, but also as a supplementary data source to improve the classification accuracy of optical data.