Autonomous clock and ephemeris predictions over extended periods have always been of interest to the Global Navigation Satellite Systems (GNSS) community, especially since the development of low-cost and low-complexity receivers operating on reduced power consumption and Location-Based Services (LBS). This paper presents a Long-Term Ephemeris (LTE) Data design concept which consists in the transmission of clock and ephemeris data for the long-term satellite orbit and clock prediction. The LTE data is computed by the ground segment and transmitted in the downlink of the GNSS satellites. Upon reception of the LTE data the GNSS receiver has all the necessary navigation data to compute its position, velocity and timing (PVT) solution and decrease the time-to-first-fix (TTFF) for an extended period of 14 days. The service design implements a satellite orbit and clock fitting optimization process based on a high-fidelity force model with the scope to offer a reduced set of parameters that can be processed by the user through a low-complexity force model. Considering a GNSS satellite equipped with a passive hydrogen maser clock (PHM) the estimated signal-in-space error (SISE) is less than 1.7 m on day 1 and 24 m on day 14 in 95% of all cases over a half-year period. The orbit and clock prediction computa-tion over a 3.5-day period based on LTE data of 24 satellites can be achieved in less than 1.4 s on a mass-market low-power processor. In the absence of updated broadcast ephemerides, the presented autonomous clock and ephemeris prediction model can address the need for an accurate and fast PVT solution in a wide range of GNSS ground or space-based applications.