Abstract: In order to solve the problem of global navigation satellite system (GNSS) receiver tracking performance degradation caused by frequent short-time signal blocking under signal blockage environments (urban canyon, dense foliage), a gated recurrent unit recurrent neural networks (GRU-RNN) assisted vector tracking loop (VTL) algorithm was proposed. The GRU-RNN network is trained by using the new information sequence of the navigation filter when the satellite signal is normal. After the network training, when the satellite signal is frequently blocked for a short time, the GRU-RNN network is used to predict the pseudo code phase error and carrier frequency error of each channel, and the predicted results are used to adjust the numerically controlled oscillator (NCO) parameters to ensure the positioning of the navigation system. The simulation results show that the proposed algorithm can effectively improve the navigation and positioning accuracy and reduce the positioning error in signal challenging environments.