Recent high-resolution satellite images provide an exciting new data sources for geospatial information acquisition. This makes it possible to extract man-made objects such as roads and buildings from these satellite images. The purpose of this paper is to explore an accurate procedure for road extraction from high resolution satellite imagery such as IKONOS. Spatial registrations of a topographic map and high resolution satellite images of the area under study are performed as data pre-processing. It is done in such a way that two data sets are unified in the same object coordinate system. Some of the most new and accurate classification techniques (Supervised, unsupervised, Maxset, Piped, Fuzzy and Neural-net classification) were applied for the required task. Two experiments were applied, first using the satellite data only as an input data for the classification process and second by using the satellite data and an elevation data source (LIDAR data) as a layer in the classification process. The results of tests illustrate that the potential of applying the neural-net classification, followed in rank with the Fuzzy technique, to road extraction and updating for both cases. Also, the results indicate that the case of using satellite data only gave an accurate output but less in accuracy than their corresponding output in the case of using satellite data and LIDAR data.