Standard for Digital Orthophoto in the National Digital Orthophoto Program (NDOP) was formulated by U.S. Geological Survey (USGS). However, the standard has not considered the requirements of large-scale city orthophoto generation, and the early procedures and algorithms in the early standard were based on earlier USGS mapping operations. To provide advanced technical basis and experience needed for the revision of the Standards for National Large-scale City Digital Orthophoto in National Digital Orthophoto Program (NDOP), this paper presents a study on theories, algorithms, and methods of large-scale urban orthoimage generation. This paper will place an emphasis oil the procedures of orthorectification for relief displacement caused by buildings and terrain, i.e., DTM-based orthoimage and DBM-based orthoimage generation, and their mergence for true orthoimage generation. Other topics, such as building occlusions and compensation, data structure to model urban buildings, shadow detection and removal, the optimization of seamline for automatic mosaic, street visibility analysis, and radiometric balance of neighbor images are briefly discussed. Finally, the accuracy comparison of orthoimages generated by traditional method and the proposed method are presented.