Abstract: Objectives:  When the GNSS-A observation technology is applied in seafloor geodesy, the GNSS antenna coordinates need to be converted to the transducer position through arm length parameters and platform attitude observations. The above conversion involves not only the error propagation of GNSS positioning error, but also the propagation of arm length error and attitude measurement error, which is a typical nonlinear error propagation problem. Therefore, it is necessary to evaluate the position of the transducer.   Methods:  In this paper, both The Guide to the Expression of Uncertainty in Measurement (GUM) based on linearized error propagation and Monte Carlo method (MCM) based on nonlinear error propagation are used to evaluate the uncertainty of transducer location. Results and   Conclusions:  The results show that:(1) GUM and MCM have a good consistency on the whole, but with the increase of nonlinear strength and measurement uncertainty, the two methods will have a significant difference; (2) The uncertainty of transducer increases with the increase of arm length, and increases with the decrease of attitude measurement and arm length measurement accuracy. This means that the longer the arm length of GNSS-A combined observation, the higher the requirement for attitude measurement accuracy; (3) Attitude measurement uncertainty has the least influence on the uncertainty of transducer location, GNSS positioning uncertainty is the most, and arm length parameter is the secondary.