Abstract: A new method on GPS weak signal tracking based on dual squareroot Kalman filters and Bayesian Estimation was proposes. It uses a new model of Kalman filter to deal with code tracking and carrier tracking respectively. Bayesian estimation theory was introduced into this method to estimate the unknown navigation data, so that the bad effect brought by reverse of data bits is deeply alleviated. The simulation results show that the weak signal with carriertonoise ratio (C/No) as low as 19 dBHz can be well tracked by the proposed method.