Abstract: There are several methods usually used for parameters estimate in precise point positioning including Kalman filters,recursive least square and square root information filtering and smoothing(SRIF&SRIS).Using Kalman filter to estimate parameters for kinematic precise point positioning(PPP)post-processing has a fatal shortcoming as the positioning precision is bad before ambiguity estimates are converged.The approach of a forward-backward Kalman filter is proposed to handle this problem,which is to sequentially carry out Kalman filter twice in both forward and backward directions.The results of the first time filtering are used as the initial information for the second time filtering while the results of the second time filtering are treated as final results of kinematic PPP.This method is simple in theory and easy for programming.Both simulated dynamic positioning test using static data and airborne dynamic positioning test were carried out.The results demonstrate that the new approach can achieve equivalent positioning accuracy as other methods.