In order to overcome the faulty channels’ influence on the receiver and improve the robustness of the vector tracking loops, a faulty channel detection algorithm, which employs the innovative sequences of the navigation filter as the input vector of a support vector machine, was proposed. A trained support vector machine was used to classify the channel as faulty or not faulty. If the channel is classified as faulty, it will not be included in the navigation filter to update the navigation results. In this way, the algorithm can eliminate the bad influence of the faulty channels. The simulation results show that this method can accurately detect faulty channels, and improve the performance and robustness of the vector tracking loops. 