Although the VNA calibration theory

at present

has been much mature

measurement uncertainty inevitably exists in practical measurement

due to random errors

linearization and measurement conditions

etc.In the existing VNA measurement uncertainty evaluation algorithms

only residual systematic error is considered

with the ignorance of the influence of system's nonlinearity

random error caused by connector and cable performance and environment conditions.In this paper

four categories of measurement error are taken into account

including residual system error

system's nonlinearity

random error and environment conditions of the measurement.Thus the rational and complete model for uncertainty in measurement is put forward and the uncertainty evaluation method for complete measurement system of VNA is established.Compared with the current VNA measurement uncertainty evaluation algorithms

a more comprehensive consideration of error factors is presented in this algorithm

leading to more reliable assessment results.The formulae of the uncertainty of S-parameter measurement are deduced in term of the theory of error limit transfer

and the corresponding method to get the error parameters is given.The algorithm is applied to measurement uncertainty evaluation for the vector network analyzer Agilent 8573ES and the results are in great agreement with the technical data provided by Agilent.