This paper presents an approach to optimize orbit geometries under satellite to satellite coverage criteria. The method described herewith is applicable to any space system concept featuring inter-satellite-links, including system concepts introducing links of opportunity between heterogeneous spacecraft at different orbits not designed a priori as a constellation. The latter is the case of federated satellite systems, that have recently been proposed as open satellite constellation concepts for opportunistic sharing of data relay and on board storage services. Federated satellite systems make use of unused telecommunications capacity available in participating spacecraft at any given time. This paper describes the satellite to satellite coverage optimization problem, a proposed optimization approach, and its analytical validation. The convergence of results is shown in terms of spatial discretization, temporal discretization and simulation period. The approach is demonstrated in a LEO to LEO network scenario, where orbital parameters of a single spacecraft are optimized to supply the maximum coverage to an example set of 6 LEO spacecraft, at different inter-satellite-link slant ranges. It is shown that increasing the inter-satellite-link maximum range above 6,000 km does not lead to further coverage benefits at LEO altitude. Orbit optimization for LEO coverage is also performed on a set of three spacecraft. In both cases, the optimum solutions feature polar inclination orbits, suggesting that said orbits may be advantageous for LEO satellite-to-satellite applications. This paper ends with conclusions outlining the future work on the satellite to satellite coverage optimization.