Multipath not only degrades the sensitivity of a GPS receiver

but also becomes the dominant error resource in GPS positioning systems. Multipath mitigation algorithm is the most significant in high sensitive and high accuracy GPS receivers. In this paper

multipath signal model in code tracking loop is studied first and a novel method for mitigation of multipath is presented which employs adaptive filtering recursive least square （RLS） algorithm. The system parameters need not be estimated in adaptive filtering algorithm and multipath signal can be filtered out directly. Under noisy scenario

the RLS algorithm is the best estimate for the filter weights with minimum least-squares errors. Simulation results show that the proposed method can reduce multipath error envelope effectively. The crossing-zero bias of the discriminator can be corrected which enhances the accuracy of code tracking in GPS receiver’s delay-locked loop （DLL）. The most importance is that the RLS algorithm is recursive and convenient to implement both in hardware and software.