Remote sensing satellites are generating large amounts of (image) data at different levels of temporal, spatial and spectral resolution, with varying levels of detail, and at various stages of processing with a variety of algorithms. This bewildering array of data sources is going to mandate a new class of information retrieval systems to help end-users access the networked information resources effectively. In this paper, we propose intelligent agent based mechanisms to locate the appropriate sources of satellite data, processing algorithms, and ancillary data that may be required at various levels of processing to answer different types of temporal and spatial queries.