A distributed orbit and propagation method for use in a predicted GPS or GNSS system, which includes a predicted GPS server (PGPS Server), a source of high accuracy orbit predictions (Orbit Server), a global reference network (GRN Server) providing real-time GPS or GNSS assistance data to the PGPS Server, a predicted GPS client (PGPS Client) running on a device equipped with a GPS or AGPS chipset. In response to requests from the PGPS Client, the PGPS Server produces and disseminates an initial seed dataset consisting of current satellite orbit state vectors and orbit propagation model coefficients. This seed dataset enables the PGPS Client to locally predict and propagate satellite orbits to a desired future time. This predictive assistance in turn helps accelerate Time To First Fix (TTFF), optimize position solution calculations and improve the sensitivity of the GPS chip present on, or coupled with, the device. In contrast with other conventional predicted GPS systems that forward large volumes of predicted orbits, synthetic ephemeris or synthetic almanac data, this method optimally reduces data transfer requirements to the client, and enables the client to locally synthesize its own predicted assistance data as needed. This method also supports seamless notification of real-time satellite integrity events and seamless integration of predicted assistance data with industry standard real-time assistance data.