A method for estimating hidden channel parameters of a GNSS navigation signal received in a multipath environment includes a sequential Bayesian estimation. A movement model forms the basis of the estimation and is designed especially for dynamic channel scenarios and takes into account the speeds of change and the variable life cycles of the path. The estimator is implemented as a recursive Bayesian filter for a combined determination of positions and suppression of multipaths. The binding of the delay of the direct path is taken into consideration via position and clock parameters. The method is carried out with GNSS satellite navigation receivers, e.g. GPS and Galileo. With the method, multipath error of received GNSS navigation signals is reduced. The multipath channel includes a direct path and Nm-I echo signals and each of these paths may be considered on or off for the estimation.