In this paper, the problem of creating a network to facilitate planning of an air reconnaissance mission is investigated. A comprehensive review of the literature shows that limited work has been done on the design of such networks in a hostile environment and that little research has accounted for terrain information. Ideally, in any routing problem a sample network is used as reference to determine paths to achieve a specified goal. The objective of this paper is to propose an automated generalized procedure developed within a Geographical Information System environment, to identify the nodes and links of a network based on the topographic information of the enemy terrain. This information is made available in the form of a Triangulated Irregular Network (TIN). Inputs to the network generation process include (I) a user-specified threshold altitude value to avoid radar detection and terrain collisions, (II) a specification of the node density in the study area and (III) the minimum/maximum lengths for the links in the desired network. The generated networks are evaluated based on traditional network analysis measures and on the area covered based on a visibility analysis. The methodology is particularly useful for air mission of Unmanned Aerial Vehicles (UAVs) and robot motion planning. As an example, a case study using elevation data from the state of Colorado is presented. Copyright Springer Science+Business Media, LLC 2007Unmanned aerial vehicles, Network generation, Geographical Information Systems (GIS), Triangulated Irregular Network (TIN), Detection altitude,