Abstract: Objectives: Implementing in-field sound speed profile (SSP) measurement not only increases the cost of seafloor geodetic observation but also constrains the timeliness of various seafloor geodetic monitoring activities. Inverting SSP based on observational information is an effective approach to replacing in-field SSP measurement. Methods: Using the publicly available GNSS-A observation dataset from the Japan Coast Guard, this paper comparatively analyzes the influence of three different constraint schemes on the accuracy of inverting SSP for the Munk model, bilinear model, and self-constructed empirical model, as well as their influence on the precision of seafloor geodetic positioning. Results: The self-constructed empirical model generally has relatively higher inversion accuracy. The RMS of the difference between the model and in-field measured SSPs is about 5 m/s in shallow waters from 10 to 1000 meters and about 1 m/s in deep waters from 1000 to 1727.80 meters. The inversion accuracy in full-depth waters is relatively highest when the sea surface is moderately constrained and the seafloor gradient is loosely constrained. When parameterizing the estimation of sound speed spatial-temporal variation compensation parameters in the ray tracing positioning model, the inversion SSPs from the self-constructed empirical model have the highest relative positioning accuracy under conditions of appropriate sea surface constraints. The RMS of difference between these profiles and the in-field measured SSP positioning results are 2 mm, 2 mm, and 2.2 cm in the E, N, and U directions, respectively. Conclusions: To achieve high-precision positioning of the inverted SSP, it is necessary to consider the spatialtemporal variation of sound speed in the positioning model. It is recommended to utilize a selfconstructed empirical model with appropriate sea surface constraints and loose seafloor gradient constraints and a two-level optimization method to achieve seafloor geodetic positioning without in-field SSP.