A large amount of aerial photographs were acquired early 1 900s for various purposes. These aerial photographs provide valuable baseline data for a variety of environmental studies. To be used for analysis together with other data, these photographs need to be referenced to a ground coordinate system. Often large topographical variations in photo ground coverage demand these photographs be orthorectified using photogrammetric methods for acceptable mapping accuracy. However, camera models for these historical photographs are usually unavailable. Additionally, land-cover change over the years makes it a great challenge to collect accurate control points for calculating camera models. This study investigated the rational function model to orthorectify historical aerial photographs. The analysis results suggest that while the linear, the 2 nd order, and the 3 rd order RFMs are all able to closely approximate the rigorous frame camera model, the linear RFM is optimal in representing imaging geometries of historical photographs.