The purpose of this research has been to develop a technique by which satellite constellations in different classes of orbit may be realistically compared. Previous work on constellation design has tended to focus on minimising the number of satellites required to provide coverage of the Earth. The variations in satellite vehicle design, which result from the use of different orbits, have, in general, been neglected in such analyses. The purpose of this research is to bridge this gap between constellation design and satellite design using a Figure of Merit. This Figure of Merit incorporates the coverage value provided by the satellite constellation, measured in terms of percentage coverage time, and the overall mass of the satellites which are required to provide this coverage. The coverage value is measured against a specific requirement, which is defined geographically, and which may be weighted by the user to reflect the relative importance of different regions. This allows arbitrary, asymmetric, real-world requirements to be adequately represented. This also marks something of a departure from previous work, in that the goal of much constellation design work has been to provide un-weighted coverage of the entire globe. Simplified mass models are developed for generic communications and surveillance satellites in a variety of orbits, and are then used to calculate the Figure of Merit for individual satellites. It is shown that the best solution depends crucially upon the geographical distribution of the requirement, and other user-defined parameters, such as the minimum elevation angle which can be tolerated. It is also shown that, for certain typical requirements, the Figure of Merit correctly identifies geostationary orbit (GEO) and low Earth orbit (LEO) as having particular advantages. iii The technique of characterising the requirement geographically may also be used as a means of optimising the orbital parameters of the candidate constellations, and a preliminary description of this procedure is also provided. The Figure of Merit Technique is then applied to representative communications satellite constellations in order to demonstrate its ability to differentiate between candidate options. The Figure of Merit technique is also used to investigate the possibility of using a surveillance satellite at very low altitudes.