Satellite Laser Ranging (SLR) is essential for the geodetic parameter determination, e.g., geocenter and station coordinates, and, therefore, for long-term stable reference frame realizations. However, the orbit modeling and the quality of the parameter estimation partially depend on the background models. This study analyses the impact of static and time-variable a priori gravity field models provided by the Center for Space Research (CSR), the International Laser Ranging Service (ILRS), and the Combination Service for Time-variable Gravity Fields (COST-G) on the SLR data processing of spherical geodetic SLR satellites (LAGEOS-1/2 and LARES) at different orbital altitudes, by comparing the estimates of Earth rotation parameters, station coordinates and observation residuals. The COST-G model is further used to examine the impact of the mean pole model and the replacement of the spherical harmonic coefficients \(C_{21}/S_{21}\) according to convention provided by the International Earth Rotation Service (IERS). While for LAGEOS-1/2 SLR data processing the a priori gravity field model has only a minor impact, the lower flying LARES satellite is more sensitive to the Earth’s gravity field and requires a more sophisticated gravity field modeling, e.g., COST-G Fitted Signal Model (FSM). In order to achieve higher consistency and thus improved solutions, the same mean pole model should be used in the SLR data processing as for the generation of the used a priori gravity field model. This study confirms the high quality of the COST-G FSM and demonstrates its suitability for potential use in the ILRS operational SLR processing.